DETAILED ACTION
This office action is in response to the communication filed on March 10, 2026. Claims 1, 3-10, and 12-20 are currently pending.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 01/27/26 has been considered by the examiner.
Remarks
In response to the terminal disclaimer filed on 03/10/26 the pending double patenting rejection has been withdrawn.
Response to Arguments
Applicant's arguments filed on March 10, 2026 have been fully considered but they are not persuasive for the following reasons:
Applicant in Pages 14-16 of the Remarks argues that the claims recite specific technical mechanisms that are fundamentally incapable of being performed as mental processes, the claimed limitations involve computational operations that cannot be performed mentally or with pen and paper, the claims do not recite an abstract idea, even assuming arguendo that the claims recite an abstract idea, the claims integrate any such idea into a practical application by improving the functioning of a computer-based media search system, and even if the claims were not found to integrate an abstract idea into a practical application, the ordered combination of multi-perspective media embeddings generated from caption data via an encoder model, dual-LLM classification and filter extraction, and filter- constrained vector similarity search is non-conventional, not routine or well-understood, and amounts to significantly more than any alleged abstract idea.
Examiner respectfully disagrees.
It is important to note that the judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements (MPEP 2106.05(a)).
Independent claim 1 and similarly independent claims 10 and 19 covers several steps, such as the generate, execute to generate, and use a filter to identify steps, that recite an abstract idea within the “Mental Processes” grouping of abstract ideas, because a person can mentally or using a pen and paper perform the limitations recited in said steps, which are discussed in detail in the current 101 rejection below.
The remaining steps in the claims that are identified as reciting additional elements, are only adding insignificant extra-solution activity to the judicial exception, and are recognized as a well understood, routine, and conventional activity within the field of computer functions, which is not sufficient to amount to significantly more than the judicial exception and are not directed to any specific improvement in computer technology.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Applicant in Pages 17-22 of the Remarks argues that Jha, Aggarwal, Myers, and Brown do not teach or even suggest the amended features “a vector store comprising a set of embeddings for a set of media items, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings, each of the multiple embeddings representing a structured data representation of a respective media perspective of the media item, wherein the structured data representations are generated based on caption data of the set of media items”, “an indexer service comprising functionality to execute an encoder model on the structured data representations to generate the set of embeddings stored in the vector store”, “a query classification service comprising functionality to: generate a first prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format, and execute a first large language model using the first prompt to generate a classification object representing classification of the query string in the structured classification format”, “a filter extraction service executing on the computer processor and comprising functionality to: generate a second prompt comprising the search request, a set of filter criteria, and definition of a structured filter format, and execute a second large language model using the second prompt to generate a filter object comprising a set of filters inferred for the query string in the structured filter format”, “use the filter object to identify a constrained set of candidate embeddings of the vector store”, as recited in amended independent claim 1 and similarly recited in amended independent claims 10 and 19.
Applicant in Pages 20-22 of the Remarks further argues that the purported combination of four references to arrive at the claimed invention reflects impermissible hindsight, the office action lacks the motivation to combine the cited prior art, and the cited prior art are not analogous art.
Examiner respectfully disagrees.
In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986).
In response to applicant's argument that the examiner's conclusion of obviousness is based upon improper hindsight reasoning, it must be recognized that any judgment on obviousness is in a sense necessarily a reconstruction based upon hindsight reasoning. But so long as it takes into account only knowledge which was within the level of ordinary skill at the time the claimed invention was made, and does not include knowledge gleaned only from the applicant's disclosure, such a reconstruction is proper. See In re McLaughlin, 443 F.2d 1392, 170 USPQ 209 (CCPA 1971).
In response to applicant’s argument that there is no teaching, suggestion, or motivation to combine the references, the examiner recognizes that obviousness may be established by combining or modifying the teachings of the prior art to produce the claimed invention where there is some teaching, suggestion, or motivation to do so found either in the references themselves or in the knowledge generally available to one of ordinary skill in the art. See In re Fine, 837 F.2d 1071, 5 USPQ2d 1596 (Fed. Cir. 1988), In re Jones, 958 F.2d 347, 21 USPQ2d 1941 (Fed. Cir. 1992), and KSR International Co. v. Teleflex, Inc., 550 U.S. 398, 82 USPQ2d 1385 (2007).
In response to applicant's argument that Brown is nonanalogous art, it has been held that a prior art reference must either be in the field of the inventor’s endeavor or, if not, then be reasonably pertinent to the particular problem with which the inventor was concerned, in order to be relied upon as a basis for rejection of the claimed invention. See In re Oetiker, 977 F.2d 1443, 24 USPQ2d 1443 (Fed. Cir. 1992).
Jha in [0015] and [0037] discloses storing structure of format of a vector embedding message, storing store at least one vector database having a plurality of vectors, each vector having an associated classification or category.
Jha in [0046]-[0049] discloses generate vector embedding message, having a response concatenation, obtain a search vector from the response vector message, using the search vector to perform a vector search of the vector database and obtain one or more associated classifications or categories, filter and rank the associated classifications or categories to output a personalized search result and result ranking.
Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development.
Therefore, Jha discloses a vector store comprising a set of embeddings for a set of media items, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings, each of the multiple embeddings representing a structured data representation of a respective media perspective of the media item, wherein the structured data representations are generated based on caption data of the set of media items.
Jha in [0084] and [0094] discloses a vector generating, an encoding generator, and a filtering and ranking generator, using an embedding model the vector generator transforms a response concatenation into a vector representation, vector generator generating vector embedding.
Jha in [0087] and [0109] discloses structure, format, or for of queries stored and indexed in database, vector embedding messages also stored and indexed in the database.
Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development.
Therefore, Jha discloses an indexer service comprising functionality to execute an encoder model on the structured data representations to generate the set of embeddings stored in the vector store.
Jha in [0012] and [0147] discloses a computer system comprising a query execution service.
Therefore, Jha discloses a query execution service.
Jha in [0012] and [0079] discloses receiving a query from a user, transmitting search results to the user, user submits a query through a client device and search results are displaying in a web browser.
Therefore, Jha discloses receiving a search request comprising a query from a client application.
Jha does not explicitly disclose receive a search request comprising a query string, but the Aggarwal reference discloses the feature.
Jha in [0010] discloses providing relevant classifications corresponding to content of a search query.
Therefore, Jha discloses a query classification service.
Jha in [0014] and [0102] discloses generating a query as a first prompt or directive to an artificial intelligence server to generate a response.
Therefore, Jha discloses generate a first prompt comprising the search request.
Jha does not explicitly disclose prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format, but the Aggarwal reference discloses the feature.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses execute a first large language model using the first prompt to generate a classification object representing classification of the query in the classification format.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses a filter extraction service executing on the computer processor and comprising functionality.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses generate a second prompt comprising the search request, a set of filter criteria, and definition of a filter format.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses execute a second large language model using the second prompt to generate a filter object comprising a set of filters inferred for the query in the filter format.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses a recaller service comprising functionality to: generate a query vector for the search request using the query the classification object, and the filter object.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses use the filter object to identify a constrained set of candidate embeddings of a vector store.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings.
Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search.
Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user.
Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query.
Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications.
Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results.
Therefore, Jha discloses provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings.
Jha discloses generating prompts comprising a search request, however, Jha does not explicitly disclose:
receive a search request comprising a query string…;
…prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format;
Aggarwal in [0054] and in Figure 2 discloses a prompt instructs a LLM of one or more tasks to be performed, prompt comprising multiple portions, such as a perspective portion, body portion, structured data portion, initialization portion.
Aggarwal in [0056] discloses perspective portion defines vocabulary or other preferences with certain attributes.
Aggarwal in [0057] discloses body portion defines searching criteria to filter search results.
Aggarwal in [0058] discloses structured data portion includes search queries, filter types, and filter values.
Aggarwal in [0059] discloses initialization portion instructs LLM to perform task described in prompt.
Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query.
Therefore, Aggarwal discloses receiving a search request comprising a query string and a prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format.
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha and Aggarwal, to have combined Jha and Aggarwal. The motivation to combine Jha and Aggarwal would be to shift the burden of query formation from a user to a natural language processing engine by using a large language model that identifies filter types and/or filter values to include in a search query based on the user’s natural language input.
Jha and Aggarwal discloses augmented search, however, Jha and Aggarwal do not explicitly disclose:
augmented semantic search;
Myers in [0070] and [0092] discloses transforming data into numerical vectors that capture semantic meaning and relationships, vectorized representations used to encode semantic information that facilitates advances retrieval techniques such as retrieval-augmented generation (RAG).
Myers in [0120] and [0121] discloses using semantic search techniques through natural language processing to understand the meaning and context of explanations and outcomes of AI model, extract semantic meaning of vector and understand the vector’s context within data, identifying similarities and differences between sets of vector representations.
Therefore, Myers discloses augmented semantic search.
It would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha, Aggarwal, and Myers, to have combined Jha, Aggarwal, and Myers. The motivation to combine Jha, Aggarwal, and Myers would be to facilitate advanced information retrieval using semantic information.
Jha, Aggarwal, and Myers disclose set of media items, however, Jha, Aggarwal, and Myers do not explicitly disclose:
wherein each of the set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications.
Brown in [0055] and [0110] discloses media is processed, analyzing video to generate detailed, time-coded descriptions of visual elements, character actions, and scene dynamics, materials include scene footage, interviews with cast and crew.
Brown in [0060], [0118], and [0122] discloses trained LLMs can summarize content, textual content from materials integrated with enriched dataset to provide comprehensive understanding of narrative, character backstories, thematic elements, and production context, fine tune a LLM specific to media.
Therefore, Brown discloses wherein each of a set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications.
It would have been obvious to a person of ordinary skill in the art, before the effective filing of the claimed invention, having the teachings of Jha, Aggarwal, Myers, and Brown, to have combined Jha, Aggarwal, Myers, and Brown. The motivation to combine Jha, Aggarwal, Myers, and Brown would be to enable a user to interact with a chatbot in a variety of ways, including text, audio, and video, by creating a multimodal chatbot using a number of AI models.
For the above reasons, Examiner states that rejection of the current Office action is proper.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1, 3-10, and 12-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
At step 1:
Independent claims 1, 10, and 19 respectively recite a system, a method, and a non-transitory computer-readable storage medium, which are directed to a statutory category such as a process, machine, or an article of manufacture.
At step 2A, prong one:
Independent claim 1 recites the limitations:
“…wherein the structured data representations are generated based on caption data of the set of media items”;
A person can mentally or using a pen and paper generate structured data representations based on caption data of a set of media items.
“…execute…on the structured data representations to generate the set of embeddings stored in the data store”;
A person can mentally or using a pen and paper execute a process on structured data representation to generate a set of embeddings stored in a data store.
“generate a first prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format”;
A person can mentally or using a pen and paper generate a first prompt comprising a search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format.
“execute…using the first prompt to generate a classification object representing classification of the query string in the structured classification format”;
A person can mentally or using a pen and paper execute a first prompt to mentally or using a pen and paper generate a classification object representing classification of a query string in a structured classification format.
“generate a second prompt comprising the search request, a set of filter criteria, and definition of a structured filter format”;
A person can mentally or using a pen and paper generate a second prompt comprising a search request, a set of filter criteria, and definition of a structured filter format.
“execute…using the first prompt to generate a classification object representing classification of the query string in the structured classification format”;
A person can mentally or using a pen and paper execute a second prompt to mentally or using a pen and paper generate a filter object comprising a set of filters inferred for a query string in a structured filter format.
“generate a query vector for the search request using the query string, the classification object, and the filter object”;
A person can mentally or using a pen and paper generate a query vector for a search request using a query string, a classification object, and a filter object.
“use the filter object to identify a constrained set of candidate embeddings of the vector store”;
A person can mentally or using a pen and paper use a filter object to identify a constrained set of candidate embeddings of a vector store.
“execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings”;
A person can mentally or using a pen and paper execute a vector similarity operation on a query vector and a constrained set of candidate embeddings to mentally or using a pen and paper generate a match set of embeddings.
The limitation, as recited above, is a process that, under its broadest reasonable
interpretation, cover steps that can be performed in the human mind or by a human
using a pen and paper, but for recitation of generic computer components.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
Independent claim 10 and similarly independent claim 19 recites the limitations:
“the structured data representations are generated based on caption data of the set of media items, wherein each media item of the set of media items is associated with multiple structured data representations of different media perspectives of the media item”;
A person can mentally or using a pen and paper generate structured data representations based on caption data of a set of media items.
“generate the set of embeddings stored in the data store”;
A person can mentally or using a pen and paper generate a set of embeddings stored in a data store.
“generate a first prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format”;
A person can mentally or using a pen and paper generate a first prompt comprising a search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format.
“execute…using the first prompt to generate a classification object representing classification of the query string in the structured classification format”;
A person can mentally or using a pen and paper execute a first prompt to mentally or using a pen and paper generate a classification object representing classification of a query string in a structured classification format.
“generate a second prompt comprising the search request, a set of filter criteria, and definition of a structured filter format”;
A person can mentally or using a pen and paper generate a second prompt comprising a search request, a set of filter criteria, and definition of a structured filter format.
“execute…using the first prompt to generate a classification object representing classification of the query string in the structured classification format”;
A person can mentally or using a pen and paper execute a second prompt to mentally or using a pen and paper generate a filter object comprising a set of filters inferred for a query string in a structured filter format.
“generate a query vector for the search request using the query string, the classification object, and the filter object”;
A person can mentally or using a pen and paper generate a query vector for a search request using a query string, a classification object, and a filter object.
“use the filter object to identify a constrained set of candidate embeddings of the vector store”;
A person can mentally or using a pen and paper use a filter object to identify a constrained set of candidate embeddings of a vector store.
“execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings”;
A person can mentally or using a pen and paper execute a vector similarity operation on a query vector and a constrained set of candidate embeddings to mentally or using a pen and paper generate a match set of embeddings.
The limitation, as recited above, is a process that, under its broadest reasonable
interpretation, cover steps that can be performed in the human mind or by a human
using a pen and paper, but for recitation of generic computer components.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
At step 2A, prong two:
This judicial exception is not integrated into a practical application.
Independent claim 1 recites the limitations:
“receive a search request comprising a query string from a client application”, which is a step of receiving data. The step is recited at a high level of generality, and amounts to mere data outputting, which is a form of insignificant extra-solution activity (MPEP 2106.05(g)).
“provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings”, which is a step of providing data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity (MPEP 2106.05(g)).
The additional elements “a system for augmented semantic search, comprising: a computer processor”, “a vector store comprising a set of embeddings for a set of media items, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings, each of the multiple embeddings representing a structured data representation of a respective media perspective of the media item”, “an indexer service comprising functionality to execute an encoder model on the structured data representations to”, “a query execution service comprising functionality to:”, “from a client application”, “a query classification service comprising functionality to:”, “execute a first large language model”, “a filter extraction service executing on the computer processor and comprising functionality to:”, “execute a second large language model”, “a recaller service comprising functionality to:”, and “of a vector store” in the steps in claim 1 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not
integrate the abstract idea into a practical application, even viewing the claims a whole,
because it does not impose any meaningful limits on practicing the abstract idea.
Independent claim 10 and similarly independent claim 19 recites the limitation:
“obtaining a set of structured data representations of media perspectives for a set of media items, the structured data representations generated based on caption data of the set of media items, wherein each media item of the set of media items is associated with multiple structured data representations of different media perspectives of the media item”, which is a step of obtaining or retrieving data. The step is recited at a high level of generality, and amounts to mere data outputting, which is a form of insignificant extra-solution activity (MPEP 2106.05(g)).
“receive a search request comprising a query string from a client application”, which is a step of receiving data. The step is recited at a high level of generality, and amounts to mere data outputting, which is a form of insignificant extra-solution activity (MPEP 2106.05(g)).
“provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings”, which is a step of providing data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity (MPEP 2106.05(g)).
The additional elements “executing an encoder model on the set of structured data representations to”, “wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings”, “from a client application”, “executing, by a computer processor, a first large language model”, “executing a second large language model”, and “of a vector store” in the steps in claim 10 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
The additional elements “a non-transitory computer-readable storage medium comprising a plurality of instructions for augmented semantic search, the plurality of instructions configured to execute on at least one computer processor to enable the at least one computer processor to:”, “execute an encoder model on the set of structured data representations to”, “wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings” “from a client application”, “execute a first large language model”, “execute a second large language model”, and “of a vector store” in the steps in claim 19 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not
integrate the abstract idea into a practical application, even viewing the claims a whole,
because it does not impose any meaningful limits on practicing the abstract idea.
At step 2B:
Independent claims 1, 10, and 19 and dependent claim 20 recites the same additional elements as identified in step 2A prong two above. These additional elements are not sufficient to amount to significantly more than the judicial exception.
Independent claim 1 recites the limitations:
“receive a search request comprising a query string from a client application”, which is a step of receiving data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of receiving or transmitting data over a network (MPEP 2106.05(d)(II)(i)).
“provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings”, which is a step of providing data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Accordingly, the additional limitations are not sufficient to amount to significantly more than the judicial exception. Therefore, the claims are directed to an abstract idea and are not patent eligible.
Independent claim 10 and similarly independent claim 19 recites the limitation:
“obtaining a set of structured data representations of media perspectives for a set of media items, the structured data representations generated based on caption data of the set of media items, wherein each media item of the set of media items is associated with multiple structured data representations of different media perspectives of the media item”, which is a step of obtaining or retrieving data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
“receive a search request comprising a query string from a client application”, which is a step of receiving data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of receiving or transmitting data over a network (MPEP 2106.05(d)(II)(i)).
“provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings”, which is a step of providing data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Dependent claim 2 and similarly dependent claim 11 recites additional limitations, such as:
“obtain a set of structured data representations of media perspectives for a set of media items, wherein the set of structured data representations is generated based on caption data of the set of media items”, which is a step of obtaining data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
“execute…on the set of structured data representations to generate the set of embeddings stored in the vector store”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper execute a set of structured data representations to mentally or using a pen and paper generate a set of embeddings stored in a vector store, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
The additional elements “an indexer service comprising functionality to:”, “execute an encode model”, and “stored in the vector store” in the steps in claim 2 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 3 and similarly dependent claims 12 and 20 recites additional limitations, such as:
“analyze the query string to determine that it is associated with two or more disjoint classifications”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1, 10, and 19, because a person can mentally or using a pen and paper analyze a query string to determine that it is associated with two or more disjoint classifications, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“asynchronously analyze the query string for each classification using separate instances…”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper analyze a query string asynchronously for each classification using separate instances, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“merge the results…to provide a comprehensive classification object”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper merge results to provide a comprehensive classification object, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
The additional elements “wherein the query classification service is further configured to:”, “of the first large language model”, and “using a multi-classification analyzer service” in the steps in claim 3 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 4 and similarly dependent claim 13 recites additional limitations, such as:
“receive multiple classification objects comprising the classification object from separate instances of the first large language model”, which is a step of receiving data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of receiving or transmitting data over a network (MPEP 2106.05(d)(II)(i)).
“analyze semantic relationships between the multiple classification objects”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper analyze semantic relationships between multiple classification objects, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“generate a unified classification hierarchy that incorporates all identified classifications”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper generate a unified classification hierarchy that incorporates all identified classifications, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“assign confidence scores to each classification within the unified hierarchy”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper assign confidence scores to each classification within a unified hierarchy, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“produce a comprehensive classification object that comprises: the unified classification hierarchy, confidence scores for each classification, potential conflicts or overlaps between classifications, and a set of suggested query refinements based on the multiple classifications”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper produce a comprehensive classification object that comprises: a unified classification hierarchy, confidence scores for each classification, potential conflicts or overlaps between classifications, and a set of suggested query refinements based on multiple classifications, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
The additional elements “wherein the multi-classification analyzer service is further configured to:”, and “of the first large language model” in the steps in claim 4 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 5 and similarly dependent claim 14 recites additional limitations, such as:
“wherein each of the set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications”, which is a step of providing data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 6 and similarly dependent claim 15 recites additional limitations, such as:
“wherein the classification object represents a classification of the entire query string, and the filter object comprises a set of filters inferred for specific entities or attributes within the query string”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper generate a classification object that represents a classification of an entire query string, and the person can mentally or using a pen and paper generate a filter object that comprises a set of filters inferred for specific entities or attributes within a query string, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 7 and similarly dependent claim 16 recites additional limitations, such as:
“generate a re-ranking prompt comprising the query string, the classification object, the filter object, and the match set of embeddings”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper generate a re-ranking prompt comprising a query string, a classification object, a filter object, and a match set of embeddings, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“execute…using the re-ranking prompt to”:
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper execute using the re-ranking prompt, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“analyze the relevance of each embedding in the match set to the query intent derived from the classification object”,
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper analyze a relevance of each embedding in a match set to a query intent derived from a classification object, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“evaluate the adherence of each embedding to the constraints specified in the filter object”,
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper evaluate adherence of each embedding to constraints specified in a filter object, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“generate a contextual importance score for each embedding based on the classification and filter analysis”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper generate a contextual importance score for each embedding based on a classification and filter analysis, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“re-rank the match set of embeddings based on the contextual importance scores”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper re-rank a match set of embeddings based on contextual importance scores, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“and provide the re-ranked match set of embeddings in response to the search request”, which is a step of providing data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
The additional elements “an intelligent re-ranking system configured to:” in the steps in claim 7 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 8 and similarly dependent claim 17 recites additional limitations, such as:
“identify named entities in the query string”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper identify named entities in a query string, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“map the named entities to predefined filter criteria”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper map a named entities to predefined filter criteria, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“and include entity-specific filter parameters in the filter object”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper include entity-specific filter parameters in a filter object, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
The additional elements “wherein the filter extraction service is further configured to:” in the steps in claim 8 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 9 and similarly dependent claim 18 recites additional limitations, such as:
“analyze the classification object and filter object for consistency with historical search patterns”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper analyze a classification object and filter object for consistency with historical search patterns, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“and adjust the classification object or filter object if inconsistencies are detected, thereby improving the accuracy of query vector generation”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 10, because a person can mentally or using a pen and paper adjust a classification object or filter object if inconsistencies are detected to improve accuracy of query vector generation, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
The additional elements “a data validation and quality control module configured to:” in the steps in claim 9 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Accordingly, dependent claims 2-9, 11-18, and 20 are also directed to abstract idea without significantly more and are not patent eligible.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 9 and 18 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 9 and 18 recites the limitation "the accuracy of query vector generation". There is insufficient antecedent basis for this limitation in the claim.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-4, 6-13, and 15-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over in Jha (US Pub 2025/0103602) in view of Aggarwal (US Pub 2025/0272317) and in further view of Myers (US Pub 2024/0411896).
With respect to claim 1, Jha discloses a system for augmented…search (Jha in [0012] and [0098] discloses a computer system for augmented search), comprising:
a computer processor (Jha in [0012] discloses a computer system comprising one or more processors);
a vector store comprising a set of embeddings for a set of media items, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings, each of the multiple embeddings representing a structured data representation of a respective media perspective of the media item, wherein the structured data representations are generated based on caption data of the set of media items (Jha in [0015] and [0037] discloses storing structure of format of a vector embedding message, storing store at least one vector database having a plurality of vectors, each vector having an associated classification or category; Jha in [0046]-[0049] discloses generate vector embedding message, having a response concatenation, obtain a search vector from the response vector message, using the search vector to perform a vector search of the vector database and obtain one or more associated classifications or categories, filter and rank the associated classifications or categories to output a personalized search result and result ranking; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
an indexer service comprising functionality to execute an encoder model on the structured data representations to generate the set of embeddings stored in the vector store (Jha in [0084] and [0094] discloses a vector generating, an encoding generator, and a filtering and ranking generator, using an embedding model the vector generator transforms a response concatenation into a vector representation, vector generator generating vector embedding; Jha in [0087] and [0109] discloses structure, format, or for of queries stored and indexed in database, vector embedding messages also stored and indexed in the database; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
a query execution service (Jha in [0012] and [0147] discloses a computer system comprising a query execution service) comprising functionality to:
receive a search request comprising a query…from a client application (Jha in [0012] and [0079] discloses receiving a query from a user, transmitting search results to the user, user submits a query through a client device and search results are displaying in a web browser; here Jha does not explicitly disclose receive a search request comprising a query string, but the Aggarwal reference discloses the feature, as discussed below);
a query classification service (Jha in [0010] discloses providing relevant classifications corresponding to content of a search query) comprising functionality to:
generate a first prompt comprising the search request…(Jha in [0014] and [0102] discloses generating a query as a first prompt or directive to an artificial intelligence server to generate a response; here Jha does not explicitly disclose prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format, but the Aggarwal reference discloses the feature, as discussed below); and
execute a first large language model using the first prompt to generate a classification object representing classification of the query…in the… classification format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
a filter extraction service executing on the computer processor and comprising functionality (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results)to:
generate a second prompt comprising the search request, a set of filter criteria, and definition of a…filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
execute a second large language model using the second prompt to generate a filter object comprising a set of filters inferred for the query…in the… filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
a recaller service comprising functionality to:
generate a query vector for the search request using the query…the classification object, and the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
use the filter object to identify a constrained set of candidate embeddings of a vector store (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results).
Jha discloses generating prompts comprising a search request, however, Jha does not explicitly disclose:
receive a search request comprising a query string…;
…prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format;
The Aggarwal reference discloses receiving a search request comprising a query string and a prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format (Aggarwal in [0054] and in Figure 2 discloses a prompt instructs a LLM of one or more tasks to be performed, prompt comprising multiple portions, such as a perspective portion, body portion, structured data portion, initialization portion; Aggarwal in [0056] discloses perspective portion defines vocabulary or other preferences with certain attributes; Aggarwal in [0057] discloses body portion defines searching criteria to filter search results; Aggarwal in [0058] discloses structured data portion includes search queries, filter types, and filter values; Aggarwal in [0059] discloses initialization portion instructs LLM to perform task described in prompt; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha and Aggarwal, to have combined Jha and Aggarwal. The motivation to combine Jha and Aggarwal would be to shift the burden of query formation from a user to a natural language processing engine by using a large language model that identifies filter types and/or filter values to include in a search query based on the user’s natural language input (Aggarwal: [0022] and [0024]).
Jha and Aggarwal discloses augmented search, however, Jha and Aggarwal do not explicitly disclose:
augmented semantic search;
The Myers reference discloses augmented semantic search (Myers in [0070] and [0092] discloses transforming data into numerical vectors that capture semantic meaning and relationships, vectorized representations used to encode semantic information that facilitates advances retrieval techniques such as retrieval-augmented generation (RAG); Myers in [0120] and [0121] discloses using semantic search techniques through natural language processing to understand the meaning and context of explanations and outcomes of AI model, extract semantic meaning of vector and understand the vector’s context within data, identifying similarities and differences between sets of vector representations).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha, Aggarwal, and Myers, to have combined Jha, Aggarwal, and Myers. The motivation to combine Jha, Aggarwal, and Myers would be to facilitate advanced information retrieval using semantic information (Myers: [0092]).
With respect to claim 2, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, further comprising:
an indexer service comprising functionality to:
obtain a set of structured data representations of media perspectives for a set of media items, wherein the set of structured data representations is generated based on caption data of the set of media items (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0065] and [0070] discloses information can be in text, image, audio, video or other computer ingestible format, information that are not text are transformed into text, video processed to extract textual information, such as extracting captions, which are analyzed and structured for the AI model, encoding text into vectors using embeddings and/or encoding; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
execute an encoder model on the set of structured data representations to generate the set of embeddings stored in the vector store (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0065] and [0070] discloses information can be in text, image, audio, video or other computer ingestible format, information that are not text are transformed into text, video processed to extract textual information, such as extracting captions, which are analyzed and structured for the AI model, encoding text into vectors using embeddings and/or encoding; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 3, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, wherein the query classification service is further configured to:
analyze the query string to determine that it is associated with two or more disjoint classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
asynchronously analyze the query string for each classification using separate instances of the first large language model (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
merge the results using a multi-classification analyzer service to provide a comprehensive classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 4, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 3, wherein the multi-classification analyzer service is further configured to:
receive multiple classification objects comprising the classification object from separate instances of the first large language model (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
analyze semantic relationships between the multiple classification objects (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
generate a unified classification hierarchy that incorporates all identified classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
assign confidence scores to each classification within the unified hierarchy (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
produce a comprehensive classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score) that comprises:
the unified classification hierarchy (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
confidence scores for each classification (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
potential conflicts or overlaps between classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score), and
a set of suggested query refinements based on the multiple classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 6, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, wherein the classification object represents a classification of the entire query string, and the filter object comprises a set of filters inferred for specific entities or attributes within the query string.
With respect to claim 7, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, further comprising an intelligent re-ranking system configured to:
generate a re-ranking prompt comprising the query string, the classification object, the filter object, and the match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
execute a third large language model using the re-ranking prompt (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score) to:
analyze the relevance of each embedding in the match set to the query intent derived from the classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
evaluate the adherence of each embedding to the constraints specified in the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score), and
generate a contextual importance score for each embedding based on the classification and filter analysis (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
re-rank the match set of embeddings based on the contextual importance scores (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
provide the re-ranked match set of embeddings in response to the search request (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 8, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, wherein the filter extraction service is further configured to:
identify named entities in the query string (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
map the named entities to predefined filter criteria (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
include entity-specific filter parameters in the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 9, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1, further comprising a data validation and quality control module configured to:
analyze the classification object and filter object for consistency with historical search patterns (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
adjust the classification object or filter object if inconsistencies are detected, thereby improving the accuracy of query vector generation (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 10, Jha discloses a method for augmented…search, comprising:
obtaining a set of structured data representations of media perspectives for a set of media items, the structured data representations generated based on caption data of the set of media items, wherein each media item of the set of media items is associated with multiple structured data representations of different media perspectives of the media item (Jha in [0015] and [0037] discloses storing structure of format of a vector embedding message, storing store at least one vector database having a plurality of vectors, each vector having an associated classification or category; Jha in [0046]-[0049] discloses generate vector embedding message, having a response concatenation, obtain a search vector from the response vector message, using the search vector to perform a vector search of the vector database and obtain one or more associated classifications or categories, filter and rank the associated classifications or categories to output a personalized search result and result ranking; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
executing an encoder model on the set of structured data representations to generate a set of embeddings stored in a vector store, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings (Jha in [0084] and [0094] discloses a vector generating, an encoding generator, and a filtering and ranking generator, using an embedding model the vector generator transforms a response concatenation into a vector representation, vector generator generating vector embedding; Jha in [0087] and [0109] discloses structure, format, or for of queries stored and indexed in database, vector embedding messages also stored and indexed in the database; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
receiving a search request comprising a query…from a client application (Jha in [0012] and [0079] discloses receiving a query from a user, transmitting search results to the user, user submits a query through a client device and search results are displaying in a web browser; here Jha does not explicitly disclose receive a search request comprising a query string, but the Aggarwal reference discloses the feature, as discussed below);
generating a first prompt comprising the search request…(Jha in [0014] and [0102] discloses generating a query as a first prompt or directive to an artificial intelligence server to generate a response; here Jha does not explicitly disclose prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format, but the Aggarwal reference discloses the feature, as discussed below); and
executing a first large language model using the first prompt to generate a classification object representing classification of the query…in the… classification format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
generate a second prompt comprising the search request, a set of filter criteria, and definition of a…filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
execute a second large language model using the second prompt to generate a filter object comprising a set of filters inferred for the query…in the… filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
generate a query vector for the search request using the query…the classification object, and the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
use the filter object to identify a constrained set of candidate embeddings of a vector store (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results).
Jha discloses generating prompts comprising a search request, however, Jha does not explicitly disclose:
receive a search request comprising a query string…;
…prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format;
The Aggarwal reference discloses receiving a search request comprising a query string and a prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format (Aggarwal in [0054] and in Figure 2 discloses a prompt instructs a LLM of one or more tasks to be performed, prompt comprising multiple portions, such as a perspective portion, body portion, structured data portion, initialization portion; Aggarwal in [0056] discloses perspective portion defines vocabulary or other preferences with certain attributes; Aggarwal in [0057] discloses body portion defines searching criteria to filter search results; Aggarwal in [0058] discloses structured data portion includes search queries, filter types, and filter values; Aggarwal in [0059] discloses initialization portion instructs LLM to perform task described in prompt; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha and Aggarwal, to have combined Jha and Aggarwal. The motivation to combine Jha and Aggarwal would be to shift the burden of query formation from a user to a natural language processing engine by using a large language model that identifies filter types and/or filter values to include in a search query based on the user’s natural language input (Aggarwal: [0022] and [0024]).
Jha and Aggarwal discloses augmented search, however, Jha and Aggarwal do not explicitly disclose:
augmented semantic search;
The Myers reference discloses augmented semantic search (Myers in [0070] and [0092] discloses transforming data into numerical vectors that capture semantic meaning and relationships, vectorized representations used to encode semantic information that facilitates advances retrieval techniques such as retrieval-augmented generation (RAG); Myers in [0120] and [0121] discloses using semantic search techniques through natural language processing to understand the meaning and context of explanations and outcomes of AI model, extract semantic meaning of vector and understand the vector’s context within data, identifying similarities and differences between sets of vector representations).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha, Aggarwal, and Myers, to have combined Jha, Aggarwal, and Myers. The motivation to combine Jha, Aggarwal, and Myers would be to facilitate advanced information retrieval using semantic information (Myers: [0092]).
With respect to claim 11, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, further comprising:
obtaining a set of structured data representations of media perspectives for a set of media items, wherein the set of structured data representations is generated based on caption data of the set of media items (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
executing an encoder model on the set of structured data representations to generate the set of embeddings stored in the vector store (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 12, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, further comprising:
analyzing the query string to determine that it is associated with two or more disjoint classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
asynchronously analyzing the query string for each classification using separate instances of the first large language model (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
merging the results using a multi-classification analyzer service to provide a comprehensive classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 13, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 12, further comprising:
receiving multiple classification objects comprising the classification object from separate instances of the first large language model (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
analyzing semantic relationships between the multiple classification objects (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
generating a unified classification hierarchy that incorporates all identified classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
assigning confidence scores to each classification within the unified hierarchy (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
producing a comprehensive classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score) that comprises:
the unified classification hierarchy (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
confidence scores for each classification (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
potential conflicts or overlaps between classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score), and
a set of suggested query refinements based on the multiple classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 15, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, wherein the classification object represents a classification of the entire query string, and the filter object comprises a set of filters inferred for specific entities or attributes within the query string.
With respect to claim 16, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, further comprising:
generating a re-ranking prompt comprising the query string, the classification object, the filter object, and the match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
executing a third large language model using the re-ranking prompt (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score) to:
analyze the relevance of each embedding in the match set to the query intent derived from the classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score),
evaluate the adherence of each embedding to the constraints specified in the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score), and
generate a contextual importance score for each embedding based on the classification and filter analysis (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score);
re-ranking the match set of embeddings based on the contextual importance scores (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score); and
providing the re-ranked match set of embeddings in response to the search request (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and filtering and ranking or re-ranking to generate and transmit personalized search results; Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0077] discloses ranking results based on relevance; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query; Aggarwal in [0100] discloses score filter types/values to measure accuracy of correctly set filters; Myers in [0087] and [0105] discloses assess data quality by evaluating AI model performance metrics, such as accuracy, precision, recall, and score).
With respect to claim 17, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, further comprising:
identifying named entities in the query string (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query);
mapping the named entities to predefined filter criteria (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query); and
including entity-specific filter parameters in the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
With respect to claim 18, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10, further comprising:
analyzing the classification object and filter object for consistency with historical search patterns (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query); and
adjusting the classification object or filter object if inconsistencies are detected, thereby improving the accuracy of query vector generation (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
With respect to claim 19, Jha discloses a non-transitory computer-readable storage medium comprising a plurality of instructions for augmented…search, the plurality of instructions configured to execute on at least one computer processor (Jha in [0147], [0152], and [0153] discloses a non-transitory computer-readable storage medium comprising a plurality of instructions for augmented…search, the plurality of instructions configured to execute on at least one computer processor) to enable the at least one computer processor to:
obtain a set of structured data representations of media perspectives for a set of media items, the structured data representations generated based on caption data of the set of media items, wherein each media item of the set of media items is associated with multiple structured data representations of different media perspectives of the media item (Jha in [0015] and [0037] discloses storing structure of format of a vector embedding message, storing store at least one vector database having a plurality of vectors, each vector having an associated classification or category; Jha in [0046]-[0049] discloses generate vector embedding message, having a response concatenation, obtain a search vector from the response vector message, using the search vector to perform a vector search of the vector database and obtain one or more associated classifications or categories, filter and rank the associated classifications or categories to output a personalized search result and result ranking; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
execute an encoder model on the set of structured data representations to generate a set of embeddings stored in a vector store, wherein each media item of the set of media items is associated with multiple embeddings of the set of embeddings (Jha in [0084] and [0094] discloses a vector generating, an encoding generator, and a filtering and ranking generator, using an embedding model the vector generator transforms a response concatenation into a vector representation, vector generator generating vector embedding; Jha in [0087] and [0109] discloses structure, format, or for of queries stored and indexed in database, vector embedding messages also stored and indexed in the database; Jha in [0078], [0135], and [0136] discloses communicating data, voice, or multimedia for user input, selection, evaluation, reporting, media provision, and so on, queries available for searching, filtering, matching, and ranking, such as searching for personalized information, such as book or music, queries indexed in memory and/or database for new categories, classifications, keywords, industries etc., queries developed and deployed for selected search topic for personalization of search results, customized filtering, and personalization of search result ranking, new fields include online music download development);
receive a search request comprising a query…from a client application (Jha in [0012] and [0079] discloses receiving a query from a user, transmitting search results to the user, user submits a query through a client device and search results are displaying in a web browser; here Jha does not explicitly disclose receive a search request comprising a query string, but the Aggarwal reference discloses the feature, as discussed below);
a query classification service (Jha in [0010] discloses providing relevant classifications corresponding to content of a search query) comprising functionality to:
generate a first prompt comprising the search request…(Jha in [0014] and [0102] discloses generating a query as a first prompt or directive to an artificial intelligence server to generate a response; here Jha does not explicitly disclose prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format, but the Aggarwal reference discloses the feature, as discussed below); and
execute a first large language model using the first prompt to generate a classification object representing classification of the query…in the… classification format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
a filter extraction service executing on the computer processor and comprising functionality (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results)to:
generate a second prompt comprising the search request, a set of filter criteria, and definition of a…filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
execute a second large language model using the second prompt to generate a filter object comprising a set of filters inferred for the query…in the… filter format (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
a recaller service comprising functionality to:
generate a query vector for the search request using the query…the classification object, and the filter object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
use the filter object to identify a constrained set of candidate embeddings of a vector store (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results);
execute a vector similarity operation on the query vector and the constrained set of candidate embeddings to generate a match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results); and
provide, in response to the search request, a result set comprising identifiers of a matching set of media items referenced by the match set of embeddings (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0056] discloses determining a confidence level or score for each classifications and combine the classifications with corresponding scores to form resulting classifications; Jha in [0098] and [0112] discloses employing retrieval augmented generation (RAG) and various machine learning models, large language models, or other supervised models trained using historical data, and filtering and ranking or re-ranking to generate and transmit personalized search results).
Jha discloses generating prompts comprising a search request, however, Jha does not explicitly disclose:
receive a search request comprising a query string…;
…prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format;
The Aggarwal reference discloses receiving a search request comprising a query string and a prompt comprising the search request, a set of categories, descriptions of the set of categories, and definition of a structured classification format (Aggarwal in [0054] and in Figure 2 discloses a prompt instructs a LLM of one or more tasks to be performed, prompt comprising multiple portions, such as a perspective portion, body portion, structured data portion, initialization portion; Aggarwal in [0056] discloses perspective portion defines vocabulary or other preferences with certain attributes; Aggarwal in [0057] discloses body portion defines searching criteria to filter search results; Aggarwal in [0058] discloses structured data portion includes search queries, filter types, and filter values; Aggarwal in [0059] discloses initialization portion instructs LLM to perform task described in prompt; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha and Aggarwal, to have combined Jha and Aggarwal. The motivation to combine Jha and Aggarwal would be to shift the burden of query formation from a user to a natural language processing engine by using a large language model that identifies filter types and/or filter values to include in a search query based on the user’s natural language input (Aggarwal: [0022] and [0024]).
Jha and Aggarwal discloses augmented search, however, Jha and Aggarwal do not explicitly disclose:
augmented semantic search;
The Myers reference discloses augmented semantic search (Myers in [0070] and [0092] discloses transforming data into numerical vectors that capture semantic meaning and relationships, vectorized representations used to encode semantic information that facilitates advances retrieval techniques such as retrieval-augmented generation (RAG); Myers in [0120] and [0121] discloses using semantic search techniques through natural language processing to understand the meaning and context of explanations and outcomes of AI model, extract semantic meaning of vector and understand the vector’s context within data, identifying similarities and differences between sets of vector representations).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Jha, Aggarwal, and Myers, to have combined Jha, Aggarwal, and Myers. The motivation to combine Jha, Aggarwal, and Myers would be to facilitate advanced information retrieval using semantic information (Myers: [0092]).
With respect to claim 20, Jha in view of Aggarwal and in further view of Myers discloses the non-transitory computer-readable storage medium of claim 19, the plurality of instructions further configured to enable the at least one computer processor to:
analyze the query string to determine that it is associated with two or more disjoint classifications (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query);
asynchronously analyze the query string for each classification using separate instances of the first large language model (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query); and
merge the results using a multi-classification analyzer service to provide a comprehensive classification object (Jha in [0020] discloses prompt to AI to perform a similarity comparison of each description associated with or corresponding to each of the categories or classifications resulting from a vector search; Jha in [0028], [0044], and [0051] discloses receive a query from user, transmit a plurality of secondary queries comprising a plurality of descriptions, words, and keywords, secondary queries comprising a prompt to AI to generate a summary and a directive to determine a similarity and difference, receive a plurality of responses to the secondary queries, transmit personalized search results and result rankings to the user; Jha in [0045]-[0050] discloses concatenate one or more responses to the secondary queries to form a concatenated response, generate vector embedding message having the response concatenation, obtain search vector from the response vector message, using the search vector to perform a search of a vector database to obtain one or more associated classifications or categories, filter and rank the classifications or categories, use the filtered and ranked information to generate and output personalized search results and rankings comprising one or more associated classifications or categories corresponding to the primary query; Jha in [0098] discloses employing retrieval augmented generation (RAG); Aggarwal in [0024] and [0047] discloses information retrieval uses entity tags to map search results to corresponding filter types and/or filter values, search query associated with a particular entity, search queries associated with entities that share common attributes; Aggarwal in [0033] discloses natural language queries associated with a structured search query provided to a second LLM using retrieval augmented generation (RAG), fine tune a second LLM, encoding LLM with filter types//values to map text to filter types and values; Aggarwal in [0057] discloses search criteria includes filter types and values, filter types are tagged entities used to search for content items; Aggarwal in [0073] discloses tokenizing unstructured query into tokens including one or more words, tokens encoded into embedding using an encoder in an embedding space, tokens associated with similar meanings positioned closed together in the embedding space; Aggarwal in [0082] and [0091] discloses LLM performs multi-class classification, classifying filter types and/or filter values associated with natural language text, perform named entity recognition which identifies and classifies natural language text as being associated with filter types/values, accepting a string as an input, such as a natural language search query, and output a string as a structured version of the natural language search query).
Claim(s) 5 and 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over in Jha (US Pub 2025/0103602) in view of Aggarwal (US Pub 2025/0272317) in view of Myers (US Pub 2024/0411896) and in further view of Brown (US Pub 2025/0265442).
With respect to claim 5, Jha in view of Aggarwal and in further view of Myers discloses the system of claim 1,
Jha, Aggarwal, and Myers disclose set of media items, however, Jha, Aggarwal, and Myers do not explicitly disclose:
wherein each of the set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications.
The Brown reference discloses wherein each of a set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications (Brown in [0055] and [0110] discloses media is processed, analyzing video to generate detailed, time-coded descriptions of visual elements, character actions, and scene dynamics, materials include scene footage, interviews with cast and crew; Brown in [0060], [0118], and [0122] discloses trained LLMs can summarize content, textual content from materials integrated with enriched dataset to provide comprehensive understanding of narrative, character backstories, thematic elements, and production context, fine tune a LLM specific to media).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing of the claimed invention, having the teachings of Jha, Aggarwal, Myers, and Brown, to have combined Jha, Aggarwal, Myers, and Brown. The motivation to combine Jha, Aggarwal, Myers, and Brown would be to enable a user to interact with a chatbot in a variety of ways, including text, audio, and video, by creating a multimodal chatbot using a number of AI models (Brown: [0020] and [0171]).
With respect to claim 14, Jha in view of Aggarwal and in further view of Myers discloses the method of claim 10,
Jha, Aggarwal, and Myers disclose set of media items, however, Jha, Aggarwal, and Myers do not explicitly disclose:
wherein each of the set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications.
The Brown reference discloses wherein each of a set of media items is associated with multiple different structured data representations of different media perspectives among the set of structured data representations, the different media perspectives comprising at least two selected from a group consisting of: hierarchical summaries, character and cast analyses, setting and time classifications, thematic and tonal analyses, plot and action detections, and micro-genre classifications (Brown in [0055] and [0110] discloses media is processed, analyzing video to generate detailed, time-coded descriptions of visual elements, character actions, and scene dynamics, materials include scene footage, interviews with cast and crew; Brown in [0060], [0118], and [0122] discloses trained LLMs can summarize content, textual content from materials integrated with enriched dataset to provide comprehensive understanding of narrative, character backstories, thematic elements, and production context, fine tune a LLM specific to media).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing of the claimed invention, having the teachings of Jha, Aggarwal, Myers, and Brown, to have combined Jha, Aggarwal, Myers, and Brown. The motivation to combine Jha, Aggarwal, Myers, and Brown would be to enable a user to interact with a chatbot in a variety of ways, including text, audio, and video, by creating a multimodal chatbot using a number of AI models (Brown: [0020] and [0171]).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to REZWANUL MAHMOOD whose telephone number is (571)272-5625. The examiner can normally be reached M-F 9-5:30.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Ann J. Lo can be reached at 571-272-9767. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/R.M/Examiner, Art Unit 2159
/ANN J LO/Supervisory Patent Examiner, Art Unit 2159
Prosecution Insights