DETAILED ACTION
This communication is in response to the Amendments and Arguments filed on 06/22/2026.
Claims 1-20 are pending and have been examined.
All previous objections/rejections not mentioned in this Office Action have been withdrawn by the examiner.
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 06/23/2026 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Response to Arguments
Applicant's arguments filed 06/22/2026 have been fully considered.
With respect to the 101 rejections, the claims have been considered in the form as submitted, and after performing the required analysis, the claims as recited are deemed as covering performance of the limitation in the mind and/or with pen and paper but for the recitation of generic computer components. The claims remain not patent eligible as set forth in the rejection below.
Applicant’s arguments with respect to the 102 and 103 rejections have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. Please see the updated claim mappings below for further detail.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim(s) 1 and 14, the limitation(s) of receiving, generating, performing, determining, associating, assigning, determining, and updating, are processes that, under broadest reasonable interpretation, covers performance of the limitation in the mind and/or with pen and paper but for the recitation of generic computer components. More specifically, the mental process of a human reading a piece of information and writing out a numerical representation of the information, finding a similar piece of information based on a similar numerical representation, writing down a value indicative of relevance and importance for each piece of information, comparing the values to see which one is higher, and using the higher value piece of information to perform a task. The recitation of a LLM reads on a human understanding the rules for how to process and respond to queries in natural language formats. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind and/or with pen and paper but for the recitation of generic computer components, then it falls within the --Mental Processes-- grouping of abstract ideas. Accordingly, the claim(s) recite(s) an abstract idea.
This judicial exception is not integrated into a practical application because the recitation of a device, processors, and computer-readable media, in claim 14 reads to generalized computer components, based upon the claim interpretation wherein the structure is interpreted using [0090-103] in the specification. Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim(s) is/are directed to an abstract idea.
The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element of using generalized computer components to receive, generate, perform, determine, associate, assign, determine, and update, amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claim(s) is/are not patent eligible.
Regarding claim(s) 18, the limitation(s) of receive, receive, assign, assign, pass, perform, and determine, as drafted, are processes that, under broadest reasonable interpretation, covers performance of the limitation in the mind and/or with pen and paper but for the recitation of generic computer components. More specifically, the mental process of a human reading a piece of information and contextual information about it and a related piece of information, writing down a value indicative of relevance and importance for each piece of information, searching a source of documents indexed in a specific way to find the most similar documents to a particular query, where the documents include the pieces of information, and comparing the values of the different pieces of information to see which one is higher. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind and/or with pen and paper but for the recitation of generic computer components, then it falls within the --Mental Processes-- grouping of abstract ideas. Accordingly, the claim(s) recite(s) an abstract idea.
This judicial exception is not integrated into a practical application because the recitation of a system, chunking tokenization and embedding component, and dynamically prioritized similarity search component, reads to generalized computer components, based upon the claim interpretation wherein the structure is interpreted using [0090-103] in the specification. Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim(s) is/are directed to an abstract idea.
The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element of using generalized computer components to receive, receive, assign, assign, pass, perform, and determine, amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claim(s) is/are not patent eligible.
With respect to claim(s) 2, 15, and 19, the claim(s) recite(s) ((claim 19) generating, binding,) generating, and binding, which reads on a human writing out the pieces of information in a specific format and combining the information with contextual information. No additional limitations are present.
With respect to claim(s) 3, the claim(s) recite(s) trigger a continuous feeding, which reads on a human using a set of specific information to determine a response based on what is read regarding the first pieces of information. No additional limitations are present.
With respect to claim(s) 4, 5, 16, 17, and 20, the claim(s) recite(s) (claims 4, 16, 20) receive, (claim 20) forward, (claims 4 and 16) query, (claims 4 and 20) return, (claims 4 and 20) return, (claims 4, 16, 20) return, (claims 5, 17, and 20) append, (claims 5, 17, and 20) pass, (claims 5 and 17) query, and (claims 5 and 17) return, which reads on a human hearing a request for information from a person, determining to use the request with a particular search strategy, looking through indexed data sources in a specific format for pieces of information similar to the request, looking up a relevance value for the first piece of information, looking up a relevance value for the second piece of information, determining which piece of information has the better value and writing it down, writing next to it additional contextual information related to the piece of information, using the piece of information and its context to determine how to respond to the query, and writing out the response. No additional limitations are present.
With respect to claim(s) 6, 8, 10, and 12, the claim(s) recite(s) specific information used to determine what weight to assign, which reads on what kind of contextual or other related information a human should use to determine how important a piece of information is. No additional limitations are present.
With respect to claim(s) 7, the claim(s) recite(s) resetting, converging, or maintaining the weightings, which reads on a human adjusting the values based on specific guidelines. No additional limitations are present.
With respect to claim(s) 9, 11, and 13, the claim(s) recite(s) what information is used to determine which weighting is higher, which reads on what kind of contextual or other related information a human should use to determine how important a piece of information is compared to other pieces of information. No additional limitations are present.
These claims further do not remedy the judicial exception being integrated into a practical application and further fail to include additional elements that are sufficient to amount to significantly more than the judicial exception.
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, 2, 4-11, and 14-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (“A Survey on Retrieval-Augmented Text Generation for Large Language Models” arXiv:2404.10981v1, 17 Apr 2024), hereinafter Huang, in view of Leandri et al. (U.S. PG Pub No. 2015/0058307), hereinafter Leandri, and further in view of Brenner et al. (U.S. PG Pub No. 2025/0165463), hereinafter Brenner.
Regarding claims 1, 14, and 18, Huang teaches
(claim 1) A method comprising (RAG workflow (Sec. 2.1)):
receiving a first data item ((claim 18) from a data source)((claim 1) to be added to a large language model (LLM))(supplementing the LLM’s training data, i.e. to be added to a LLM, with current information from external sources, from a data source, through a retrieval model, i.e. receiving a first data item (Sec. 2 intro, 2.1, 2.1.1));
generating a first embedding for the first data item (semantic vector representations of texts, i.e. first embedding for the first data item, are generated for data collections as part of indexing (Sec 2.1, 2.1.1));
… identify an instance of the … data item that is associated with the LLM (supplementing the LLM’s training data, i.e. associated with the LLM, with current information from external sources through a retrieval model, i.e. identify an instance of the … data item (Sec. 2 intro, 2.1, 2.1.1));
determining a first weighting for the first data item (during retrieval, documents, i.e. first data item, are initially ranked according to relevance, such as a relevance score, and can also be re-ranked, i.e. determining a first weighting (Sec. 2.2.2-2.2.3));
associating the first data item with the first embedding and the first weighting (data collections of text are stored with their vectors, i.e. associating the first data item with the first embedding, where documents are ranked according to their relevance, and prior relevance evaluations are used in the re-ranking and filtering process, i.e. associating the first data item…and the first weighting (Sec. 2.2.2-2.2.3));
assigning a second weighting to the instance of the … data item (during retrieval, documents, i.e. instance of the…data item, are initially ranked according to relevance, such as a relevance score, and can also be re-ranked, i.e. assigning a second weighting (Sec. 2.2.2-2.2.3));
determining which of the first or second weightings is a higher weighting (documents are ranked and re-ranked according to their relevance, where documents may be further filtered to remove documents that fail to meet specified quality or relevance standards, such as excluding documents below a minimum relevance score threshold, i.e. determining which of the first or second weightings is a higher weighting (Sec. 2.2.2-2.2.3)); and
(claim 1) updating the LLM with … the first data item or the instance of the … data item associated with the higher weighting (the retrieved information that was not filtered out, i.e. the first data item or the instance of the … data item associated with the higher weighting, is merged with the user’s query to improve generation by the LLM, i.e. updating the LLM (Sec. 2.2.2-2.2.4, 6.1)).
(claim 14) assigning a higher search priority to one of the first data item or the instance of the … data item based on the higher weighting (documents are ranked and re-ranked, i.e. assigning a higher search priority to one of the first data item or the instance of the … data item, according to their relevance, where documents may be further filtered to remove documents that fail to meet specified quality or relevance standards, such as excluding documents below a minimum relevance score threshold, i.e. based on the higher weighting (Sec. 2.2.2-2.2.3)).
(claim 18) pass a query related to the first data item and to the instance of the … data item (the query is adjusted to be a better match with indexed data, and then search algorithms are employed to identify documents that match a user’s query, i.e. pass a query related to the first data item and to the instance of the … data item (Sec. 2.2.1-2.2.2));
(claim 18) perform a similarity search and context retrieval from one or more vectorized knowledgebases associated with the first data item and the instance of the … data item (the documents and their semantic vector representations are indexed for retrieval, i.e. retrieval from one or more vectorized knowledgebases associated with the first data item and the instance of the … data item, where the vector distance between documents and queries is measured to determine semantic similarity for retrieval and ranking, i.e. perform a similarity search, which enables dynamic retrieval of contextually relevant examples, i.e. context retrieval (Sec. 2.1.1-2.1.2, 2.2.1-2.2.2, 3.1)).
While Huang provides identifying documents similar to a query through semantic similarity of vectors between documents and queries, Huang does not specifically teach identifying an instance of a first data item based on similarity or weighting based on descriptions, and thus does not teach
performing a similarity search with the first embedding ((claim 1) to) identify an instance of the first data item….
(claim 18) receive descriptive information about the first data item and about an instance of the first data item;
(claim 18) assign a first weighting to the first data item based on the descriptive information about the first data item;
(claim 18) assign a second weighting to the instance of the first data item based on the descriptive information about the instance of the first data item.
Leandri, however, teaches performing a similarity search with the first embedding ((claim 1) to) identify an instance of the first data item … (information from different content are compared to detect the similarities of the content, i.e. performing a similarity search, and, if similarities are detected, determining that an item was copied from an information source, i.e. identify an instance of the first data item [0012-3],[0019-27]).
Where Huang specifically teaches that semantic similarity is determined by a measuring of vector distances, i.e. a similarity search with the first embedding (Sec. 2.1.2).
(claim 18) receive descriptive information about the first data item and about an instance of the first data item (a time-stamp and source, i.e. descriptive, are associated with the original content and the copy, i.e. receive…about the first data item and about an instance of the first data item [0006-23]);
(claim 18) assign a first weighting to the first data item based on the descriptive information about the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on the descriptive information about the first data item, therefore the original content would have a higher weight than a copy, i.e. assign a first weighting to the first data item [0006-23]);
(claim 18) assign a second weighting to the instance of the first data item based on the descriptive information about the instance of the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on the descriptive information about the instance of the first data item, therefore the original content would have a higher weight than a copy, i.e. assign a second weighting to the instance of the first data item [0006-23]).
Huang and Leandri are analogous art because they are from a similar field of endeavor in retrieving relevant information. Thus, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the identifying documents similar to a query through semantic similarity of vectors between documents and queries teachings of Huang with the comparison of content to identify copies of content based on similarity as taught by Leandri. It would have been obvious to combine the references to quickly provide relevant information that favors the original source of content rather than a copy (Leandri [0006-7],[0016-9]).
While Huang in view of Leandri provides merging high ranking documents with the user query for LLM generation, Huang in view of Leandri does not specifically teach updating the LLM with one of the data items that has the higher weighting, and thus does not teach
(claim 1) updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting.
Brenner, however, teaches
((claim 14) A device/(claim 18) system) comprising (device [0111]):
(claims 14 and 18) one or more processors (processing device [0111]); and
(claims 14 and 18) one or more non-transitory computer-readable media storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising (software instructions embodied in a computer readable medium and executed by the processing device [0111-2]):
(claim 1) updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting (the most relevant, or top-K, document chunks, i.e. the first data item or the instance of the first data item, where K may be equal to 1, i.e. one…associated with the higher weighting, may be provided to the generative model, i.e. updating the LLM [0050]).
Huang, Leandri, and Brenner are analogous art because they are from a similar field of endeavor in retrieving relevant information. Thus, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the merging high ranking documents with the user query for LLM generation teachings of Huang, as modified by Leandri, with the provision of only the top ranked document chunk as taught by Brenner. It would have been obvious to combine the references to enable dynamically adjusting the value of K based on a number of factors for each ranker being used (Brenner [0050]).
Regarding claims 2, 15, and 19, Huang in view of Leandri and Brenner teaches claims 1, 14, and 18, and Huang further teaches
(claim 19) generate a first embedding associated with the first data item (semantic vector representations of texts, i.e. first embedding for the first data item, are generated for data collections as part of indexing (Sec 2.1, 2.1.1));
(claim 19) bind the first embedding to the first data item with augmented metadata associated with the first weighting assigned to the first data item (data collections of text are stored with their vectors, i.e. bind the first embedding to the first data item, where documents are ranked according to their relevance, and prior relevance evaluations are used in the re-ranking and filtering process, i.e. bind…the first data item…associated with the first weighting, and where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented metadata (Sec. 2.2.1-2.2.3, 3.3));
generating a second embedding for the instance of the first data item (semantic vector representations of texts, i.e. second embedding for the instance of the first data item, are generated for data collections as part of indexing (Sec 2.1, 2.1.1)); and
binding the second embedding to the instance of the first data item with augmented metadata associated with the second weighting assigned to the instance of the first data item (data collections of text are stored with their vectors, i.e. binding the second embedding to the instance of the first data item, where documents are ranked according to their relevance, and prior relevance evaluations are used in the re-ranking and filtering process, i.e. binding… the instance of the first data item…with the second weighting, and where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented metadata (Sec. 2.2.1-2.2.3, 3.3)).
Where Leandri teaches the relationship of the original and copied content [0012-3],[0019-27].
And where the motivation to combine is the same as previously presented.
Regarding claims 4, 16, and 20, Huang in view of Leandri and Brenner teaches claims 2, 15, and 19, and Huang further teaches
receiving a query ((claims 4 and 20) where the query is applicable to the first data item and to the instance of the first data item)(the query is adjusted to be a better match with indexed data, and then search algorithms are employed to identify documents that match a user’s query, i.e. receiving a query related to the first data item and to the instance of the … data item (Sec. 2.2.1-2.2.2));
(claim 20) forward the query for a similarity search (the documents and their semantic vector representations are indexed for retrieval, where the vector distance between documents and queries is measured to determine semantic similarity for retrieval and ranking, i.e. forward the query for a similarity search, which enables dynamic retrieval of contextually relevant examples(Sec. 2.1.1-2.1.2, 2.2.1-2.2.2, 3.1));
querying one or more (claims 4 and 20) vectorized) databases for the first and second embeddings (the documents and their semantic vector representations are indexed for retrieval, i.e. one or more vectorized databases, where the vector distance between documents and queries is measured to determine semantic similarity, i.e. querying…for the first and second embeddings, for retrieval and ranking of contextually relevant examples (Sec. 2.1.1-2.1.2, 2.2.1-2.2.2, 3.1)).
(claim 20) passing query with the augmented context information to a large language model (LLM) (the retrieved information that was not filtered out, where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented context information, is merged with the user’s query to improve generation by the LLM, i.e. passing query with…to a large language model (LLM) (Sec. 2.2.1-2.2.4, 3.3, 6.1)).
Where Leandri further teaches
(claims 4 and 20) returning the first weighting assigned to the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, therefore the original content would have a higher weight than a copy, i.e. the first weighting assigned to the first data item, where the search results are hierarchized based on the weight of the source, i.e. returning the first weighting [0006-23]);
(claims 4 and 20) returning the second weighting assigned to the instance of the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, therefore the original content would have a higher weight than a copy, i.e. the second weighting assigned to the instance of the first data item, where the search results are hierarchized based on the weight of the source, i.e. returning the second weighting [0006-23]).
And Brenner further teaches returning one of the first data item or the instance of the first data item associated with the higher weighting (the most relevant, or top-K, document chunks, i.e. the first data item or the instance of the first data item, where K may be equal to 1, i.e. one…associated with the higher weighting, may be provided to the generative model, i.e. returning [0050]).
And where the motivation to combine is the same as previously presented.
Regarding claims 5 and 17, Huang in view of Leandri and Brenner teaches claims 4 and 16, and Huang further teaches
appending the query with augmented context information associated with the one of the first data item or the instance of the first data item associated with the higher weighting (the retrieved information that was not filtered out, i.e. associated with the one of the first data item or the instance of the first data item associated with the higher weighting, where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented context information, is merged with the user’s query to improve generation by the LLM, i.e. appending the query (Sec. 2.2.1-2.2.4, 3.3, 6.1));
passing the augmented context information to the LLM (the retrieved information that was not filtered out, where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented context information, is merged with the user’s query to improve generation by the LLM, i.e. passing…to the LLM (Sec. 2.2.1-2.2.4, 3.3, 6.1));
querying the LLM with the appended query (the retrieved information that was not filtered out, where the data modification for indexing can be performed using metadata to enhance the raw text, i.e. augmented context information, is merged with the user’s query to improve generation by the LLM, i.e. querying the LLM with the appended query (Sec. 2.2.1-2.2.4, 3.3, 6.1)); and
returning a response from the LLM based on the appended query (the LLM generates and outputs a coherent and relevant response, i.e. returning a response from the LLM, based on the retrieved information merged with the query, i.e. based on the appended query (Sec. 2.2.1-2.2.4, 3.3, 6.1-6.2)).
Regarding claim 6, Huang in view of Leandri and Brenner teaches claim 4, and Leandri further teaches
determining which of the first or second weightings is the higher weighting according to a least one of:
determining which of the first or second weightings is the higher weighting based on a most recent time of generation (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on a most recent time of generation, therefore the original content would have a higher weight than a copy, i.e. determining which of the first or second weightings is the higher weighting [0006-23]);
determining which of the first or second weightings is the higher weighting is based on a higher priority origin of information; or
determining which of the first or second weightings is the higher weighting based on a higher vulnerability risk.
Where the motivation to combine is the same as previously presented.
Regarding claim 7, Huang in view of Leandri and Brenner teaches claim 1, and Huang further teaches
after updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting, processing the first and second weightings according to at least one of:
resetting the first and second weightings after a determined period of time;
converging the first and second weightings into a single weighting; or
maintaining the first and second weightings until a condition is met (data collections of text are stored with their vectors, where documents are ranked according to their relevance, i.e. after updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting, processing the first and second weighting, and prior relevance evaluations are used, i.e. maintaining the first and second weightings, in the re-ranking and filtering process, i.e. until a condition is met (Sec. 2.2.2-2.2.3)).
Regarding claim 8, Huang in view of Leandri and Brenner teaches claim 1, and Leandri further teaches
determining the first weighting for the first data item includes determining the first weighting for the first data item based on a time of generation of the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on a time of generation of the first data item, therefore the original content would have a higher weight than a copy, i.e. determining the first weighting for the first data item includes determining the first weighting for the first data item [0006-23]); and
assigning the second weighting to the instance of the data item includes assigning the second weighting to the instance of the first data item based on a time of generation of the instance of the first data item (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on a time of generation of the instance of the first data item, therefore the original content would have a higher weight than a copy, i.e. assigning the second weighting to the instance of the data item includes assigning the second weighting to the instance of the first data item [0006-23]).
Where the motivation to combine is the same as previously presented.
Regarding claim 9, Huang in view of Leandri and Brenner teaches claim 8, and Leandri further teaches
determining which of the first or second weightings is the higher weighting includes determining which of the first or second weightings is based on a most recent time of generation (a higher weight is assigned to the source associated with the earliest time-stamp for the content, i.e. based on a most recent time of generation, therefore the original content would have a higher weight than a copy, i.e. determining which of the first or second weightings is the higher weighting includes determining which of the first or second weightings [0006-23]).
Where the motivation to combine is the same as previously presented.
Regarding claim 10, Huang in view of Leandri and Brenner teaches claim 1, and Leandri further teaches
determining the first weighting includes determining the first weighting based on a first origin of information describing the first data item (a higher weight is assigned to the source, such as a site or an author, associated with the earliest time-stamp for the content, which is the source from which the information is copied, i.e. based on a first origin of information describing the first data item, therefore the original content would have a higher weight than a copy, i.e. determining the first weighting includes determining the first weighting [0006-23]); and
assigning the second weighting to the instance of the first data item includes assigning the second weighting to the instance of the first data item based on a second origin of information describing the instance of the first data item (a higher weight is assigned to the source, such as a site or an author, associated with the earliest time-stamp for the content, rather than another site or author copying the content, i.e. based on a second origin of information describing the instance of the first data item, therefore the original content would have a higher weight than a copy, i.e. assigning the second weighting to the instance of the first data item includes assigning the second weighting to the instance of the first data item [0006-23]).
Where the motivation to combine is the same as previously presented.
Regarding claim 11, Huang in view of Leandri and Brenner teaches claim 10, and Leandri further teaches
determining which of the first or second weightings is a higher weighting includes determining which of the first or second weightings is based on a higher priority origin of information (a higher weight is assigned to the source, such as a site or an author, associated with the earliest time-stamp for the content, which is the source from which the information is copied, i.e. based on a higher priority origin of information, therefore the original content would have a higher weight than a copy, i.e. determining which of the first or second weightings is a higher weighting includes determining which of the first or second weightings [0006-23]).
Where the motivation to combine is the same as previously presented.
Claim(s) 3 is/are rejected under 35 U.S.C. 103 as being unpatentable over Huang, in view of Leandri, in view of Brenner, and further in view of Ge et al. (“IN-CONTEXT AUTOENCODER FOR CONTEXT COMPRESSION IN A LARGE LANGUAGE MODEL”, arXiv:2307.06945v3, 18 Mar 2024), hereinafter Ge.
Regarding claim 3, Huang in view of Leandri and Brenner teaches claim 2, and Huang further teaches
the first and second embeddings trigger a continuous feeding of augmented metadata and associated —information-- for each of the first data item and the instance of the first data item into the LLM (the retrieved documents that are relevant to the query based on semantic vector representation, i.e. the first and second embeddings…each of the first data item and the instance of the first data item, are augmented, such as with metadata, i.e. augmented metadata, where the retrieved information that passes the filtering step is merged with the user’s query for response generation by the LLM, which can involve generating multiple independed recitations and employing a vote system to determine the most appropriate answer, i.e. continuous feeding…associated information…into the LLM (Sec. 2.2.1-2.2.4, 3.3, 6.1)).
While Huang in view of Leandri and Brenner provides providing retrieved documents to the input context for the LLM, Huang in view of Leandri and Brenner does not specifically teach the format is an associated embedding, and thus does not teach
trigger a continuous feeding of …associated embeddings for each of the first data item and the instance of the first data item into the LLM.
Ge, however, teaches trigger a continuous feeding of …associated embeddings for each of the first data item and the instance of the first data item into the LLM (the long-context text to be used for context, i.e. each of the first data item and the instance of the first data item, are encoded into memory slots, i.e. associated embeddings, and provided as part of the prompt for the LLM to formulate a response, i.e. trigger a continuous feeding…into the LLM Fig. 1,2,(Sec. 1 and 2)).
Huang, Leandri, Brenner, and Ge, are analogous art because they are from a similar field of endeavor in retrieving relevant information for a user. Thus, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the providing retrieved documents to the input context for the LLM teachings of Huang, as modified by Leandri and Brenner, with the use of autoencoding context into memory slots as taught by Ge. It would have been obvious to combine the references to express the same content with a shorter context, thereby enhancing the model’s ability to handle long contexts with improved latency and memory cost during inference (Ge (Sec. 1)).
Claim(s) 12 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Huang, in view of Leandri, in view of Brenner, and further in view of Cameron et al. (U.S. PG Pub No. 2025/0165616), hereinafter Cameron.
Regarding claim 12, Huang in view of Leandri and Brenner teaches claim 1.
While Huang in view of Leandri and Brenner provides determining weightings for two pieces of information, Huang in view of Leandri and Brenner does not specifically teach weighting based on vulnerabilities, and thus does not teach
determining the first weighting includes determining the first weighting based on a first vulnerability associated with the first data item; and
assigning the second weighting to the instance of the first data item includes assigning the second weighting to the instance of the first data item based on a second vulnerability associated with the instance of the first data item.
Cameron, however, teaches determining the first weighting includes determining the first weighting based on a first vulnerability associated with the first data item (a policy may be used by a machine learning model to generate a higher weight for a threat value of the security vulnerability related to cryptography in the cloud-based platform, i.e. determining the first weighting based on a first vulnerability associated with the first data item, and a lower weight to the same vulnerability in the mobile application-based platform, where the weight is generated for input information, i.e. determining the first weighting [0058-59]); and
assigning the second weighting to the instance of the first data item includes assigning the second weighting to the instance of the first data item based on a second vulnerability associated with the instance of the first data item (a policy may be used by a machine learning model to generate a higher weight for a threat value of the security vulnerability related to cryptography in the cloud-based platform, and a lower weight to the same vulnerability in the mobile application-based platform, i.e. assigning the second weighting to the instance of the first data item based on a second vulnerability associated with the instance of the first data item, where the weight is generated for input information, i.e. assigning the second weighting to the instance of the first data item [0058-59]).
Huang, Leandri, Brenner, and Cameron are analogous art because they are from a similar field of endeavor in weighting information in a system for generating output. Thus, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the determining weightings for two pieces of information teachings of Huang, as modified by Leandri and Brenner, with having different weighting for different threat values of security vulnerabilities as taught by Cameron. It would have been obvious to combine the references to enable improved determinations of real-time security vulnerabilities with respect to different infrastructure and architecture components so that network engineers may select the best platform and reduce the chance of data breaches (Cameron [0025]).
Regarding claim 13, Huang in view of Leandri, Brenner, and Cameron teaches claim 12, and Cameron further teaches
determining which of the first or second weightings is the higher weighting includes determining which of the first or second weightings is based on a higher vulnerability risk (a policy may be used by a machine learning model to generate a higher weight for a threat value of the security vulnerability related to cryptography in the cloud-based platform, and a lower weight to the same vulnerability in the mobile application-based platform, i.e. determining which of the first or second weightings is the higher weighting, where the threat value indicates a level of risk associated with a given security vulnerability, i.e. based on a higher vulnerability risk [0058-59]).
Where the motivation to combine is the same as previously presented.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NICOLE A K SCHMIEDER whose telephone number is (571)270-1474. The examiner can normally be reached 8:00 - 5:00 M-F.
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, Pierre-Louis Desir can be reached at (571) 272-7799. 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.
/NICOLE A K SCHMIEDER/Primary Examiner, Art Unit 2659
Prosecution Insights