DETAILED ACTION
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 .
Claim(s) 1-20 is/are pending and has/have been examined.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 05/01/2024 and 08/20/2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Drawings
The drawings are objected to because of the following informalities:
Fig. 2 - element 224 not in spec;
Fig. 6 - element 418 in fig. is referred to as 618 in spec [0095].
Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. The figure or figure number of an amended drawing should not be labeled as “amended.” If a drawing figure is to be canceled, the appropriate figure must be removed from the replacement sheet, and where necessary, the remaining figures must be renumbered and appropriate changes made to the brief description of the several views of the drawings for consistency. Additional replacement sheets may be necessary to show the renumbering of the remaining figures. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Claim Objections
Claims 2, 6, 8, 10, 12, 13, and 15 are objected to because of the following informalities:
Claims 2 and 15 recite “assigning a first weighting” and “assigning a second weighting”. The Examiner suggests amending the claim(s) to recite –assigning the first weighting—and –assigning the second weighting--, respectively, in order to maintain clear antecedent basis.
Claim 6 recites “a higher weighting” in each of the three limitations. The Examiner suggests amending the claim(s) to recite –the higher weighting—in all three instances in order to maintain clear antecedent basis.
Claims 8, 10, and 12, recite “a first weighting” in the first limitation of each claim. The Examiner suggests amending the claim(s) to recite –the first weighting-- in order to maintain clear antecedent basis.
Claim 13 recites “a higher weighting”. The Examiner suggests amending the claim(s) to recite –the higher weighting— in order to maintain clear antecedent basis.
Appropriate correction is required.
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 19 and 20 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.
Claim 19 recites “the instance of the data” in the last limitation. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 is rejected as being dependent upon a rejected base claim.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: “chunking, tokenization and embedding component” and “dynamically prioritized similarity search component” in claim 18.
Regarding the terms “chunking, tokenization and embedding component” and “dynamically prioritized similarity search component”, the terms are generic placeholders. There is no evidence that one or ordinary skill in the art would understand the structure by looking at the terms. Further, the terms are modified by the functional language “operative to”, but are not modified by a sufficient structure for performing the claimed function.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
The “chunking, tokenization and embedding component” and “dynamically prioritized similarity search component” are embodied as a processor that performs the method, as per the specifications [0052].
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
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, determining, assigning, assigning, determining, and updating, 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 finding a related piece of information, 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, determine, assign, 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) generating, binding, generating, 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 and 16) receive, (claim 20) forward, (claims 4 and 16) query, (claims 4 and 16) return, (claims 4 and 16) return, (claims 4 and 16) 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 § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-5, 8-11, and 14-20 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Madisetti et al. (US PG Pub No. 2025/0190460), hereinafter Madisetti.
Regarding claims 1 and 14, Madisetti teaches
(claim 1) A method comprising (a method [0020]):
(claim 14) A device comprising (device [0211]):
(claim 14) one or more processors (a processor [0211]); and
(claim 14) 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 (computer readable storage operable to store software with commands that the processor executes [0211]):
receiving a first data item to be added to a large language model (LLM) (the user has documents that are processed through a LASER system to be refined, i.e. receiving a first data item, and fed to the SCORE-RAG system, which further sends documents and/or chunks to the LLM, i.e. to be added to a LLM [0170-2]);
determining that an instance of the first data item is present in the LLM (the SCORE-RAG system creates optimized chunks of information and associated metadata based relationship information between chunks, such as the same category of information, where chunks may be organized based on the relationship to each other and to categories, i.e. determining that an instance of the first data item is present in the LLM [0170-2]);
assigning a first weighting to the first data item to be added to the LLM (the SCORE-RAG system takes the categorized chunks, including metadata such as relationships between chunks, and performs scoring and ranking to each chunk, such as a more recent chunk may be ranked higher, i.e. assigning a first weighting to the first data item to be added to the LLM [0170-2],[0181]);
assigning a second weighting to the instance of the first data item (the SCORE-RAG system takes the categorized chunks, including metadata such as relationships between chunks, and performs scoring and ranking on each chunk, such as a more recent chunk may be ranked higher, i.e. assigning a first weighting to the first data item to be added to the LLM [0170-2],[0181]);
determining which of the first or the second weightings is a higher weighting (the chunks are scored and ranked, i.e. determining which of the first or the second weightings is a higher weighting [0170],[0172],[0181]); and
updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting (the highly scored chunks, such as the top chunk, may be sent to the LLM for processing to create results, i.e. updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting [0157],[0172],[0181]).
Regarding claim 18, Madisetti teaches
A system comprising (device [0211]):
a chunking, tokenization and embedding component operative (processor configured to execute commands received from the software, i.e. component [0211]):
to receive a first data item from a data source to be added to a large language model (LLM) (a data loader reads documents from a data source, i.e. to receive a first data item from a data source, which will be processed and sent to the LLM, i.e. to be added to a LLM [0151-2],[0170-2]);
to receive descriptive information about the first data item and about an instance of the first data item (the documents are split into chunks, i.e. the first data item and about an instance of the first data item, and converted into vector embeddings and related metadata is extracted and stored with associated split, where the metadata can be descriptive metadata about the chunk itself, i.e. to receive descriptive information about [0152],[0155],[0181],[0183],[0188],[0195-6]);
to assign a first weighting to the first data item based on the descriptive information about the first data item (the SCORE-RAG system takes the categorized chunks, including metadata such as relationships between chunks, timestamps, and citation analysis, i.e. based on the descriptive information about the first data item, and performs scoring and ranking to each chunk, such as a more recent chunk may be ranked higher, i.e. to assign a first weighting to the first data item [0170-2],[0181-3],[0195]);
to 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 (the SCORE-RAG system takes the categorized chunks, including metadata such as relationships between chunks, timestamps, and citation analysis, i.e. based on the descriptive information about the instance of the first data item, and performs scoring and ranking to each chunk, such as a more recent chunk may be ranked higher, i.e. to assign a second weighting to the instance of the first data item [0170-2],[0181-3],[0195]);
to pass a query to a dynamically prioritized similarity search component directed to the first data item and to the instance of the first data item (the system uses context-optimized retrieval techniques, such as a search query, to search a large corpus of documents and retrieve a subset of document chunks related to the user query, such as through a vector similarity search, i.e. to pass a query to a…similarity search component directed to the first data item and to the instance of the first data item, using vector databases where the document chunks are stored, and where the SCORE-RAG system scores and ranks the chunks based on relevance to a particular query, i.e. dynamically prioritized similarity search component [0152],[0156-9],[0170], [0172],[0181],[0183],[0189]);
the dynamically prioritized similarity search component being operative (processor configured to execute commands received from the software, i.e. component [0211]):
to perform a similarity search and context retrieval from one or more vectorized knowledgebases associated with the first data item and the instance of the first data item (the system uses context-optimized retrieval techniques, such as a search query, to search a large corpus of documents and retrieve a subset of document chunks related to the user query, such as through a vector similarity search, i.e. to perform a similarity search and context retrieval, using vector databases where the document chunks are stored, i.e. from one or more vectorized knowledgebases associated with the first data item and the instance of the first data item, and where the SCORE-RAG system scores and ranks the chunks based on relevance to a particular query, i.e. dynamically prioritized similarity search component [0152],[0156-9],[0170], [0172],[0181],[0183],[0189]); and
to determine which of the first or second weightings is a higher weighting (the SCORE-RAG system scores the chunks and ranks them, i.e. to determine which of the first or second weightings is a higher weighting, so that the highly scored chunks may be sent to the LLM [0170],[0172]).
Regarding claims 2, 15, and 19, Madisetti teaches claims 1, 14, and 18, and further teaches
assigning a first weighting to the first data item to be added to the LLM includes:
generating a first embedding in a first vectorized ((claims 2 and 15) database/(claim 19) knowledgebase), the first embedding associated with the first data item (documents are converted into vector embeddings and are split into smaller chunks creating document splits, i.e. first embedding associated with the first data item, which are stored in the form of vector embeddings in a vector database along with extracted metadata, i.e. generating a first embedding in a first vectorized database, where multiple specialized vector databases can be employed, i.e. a first vectorized database [0152],[0155-6],[0187],[0189]);
binding the first embedding to the first data item with augmented metadata associated with the first weighting assigned to the first data item (documents are converted into vector embeddings and are split into smaller chunks creating document splits, which are stored in the form of vector embeddings in a vector database along with extracted metadata, where the metadata is appended to the vector using a tag assigner, i.e. binding the first embedding to the first data item with augmented metadata, such as contextual information and a timestamp tracker, which can be used to score the chunk, i.e. assigning a first weighting to the first data item…includes…augmented metadata associated with the first weighting assigned to the first data item [0152],[0155-6],[0181],[0183],[0187]);
assigning a second weighting to the instance of the first data item includes:
generating a second embedding in a second vectorized ((claims 2 and 15) database/(claim 19) knowledgebase), the second embedding associated with the instance of the first data item (documents are converted into vector embeddings and are split into smaller chunks creating document splits, i.e. second embedding associated with the instance of the first data item, which are stored in the form of vector embeddings in a vector database along with extracted metadata, i.e. generating a second embedding in a second vectorized database, where multiple specialized vector databases can be employed, i.e. a second vectorized database [0152],[0155-6],[0187],[0189]); 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 (documents are converted into vector embeddings and are split into smaller chunks creating document splits, which are stored in the form of vector embeddings in a vector database along with extracted metadata, where the metadata is appended to the vector using a tag assigner, i.e. binding the second embedding to the instance of the first data item with augmented metadata, such as contextual information and a timestamp tracker, which can be used to score the chunk, i.e. assigning a second weighting to the instance of the first data item…includes…augmented metadata associated with the second weighting assigned to the instance of the first data item [0152],[0155-6],[0181],[0183],[0187]).
Regarding claim 3, Madisetti teaches claim 2, and further teaches
the first and second embeddings trigger a continuous feeding of augmented metadata and associated embeddings for each of the first data item and the instance of the first data item into the LLM (the SCORE-RAG system may perform online real-time processing to create relevant context by taking the categorized chunks from documents, which are embedded as vectors, i.e. the first and second embeddings trigger…associated embeddings for each of the first data item and the instance of the first data item, and other metadata, such as the metadata appended to the vector, i.e. augmented metadata, to send to the LLM for processing to create results for the user, i.e. trigger a continuous feeding…into the LLM [0152],[0155-6],[0170],[0172],[0181],[0183],[0187]).
Regarding claims 4, 16, and 20, Madisetti teaches claims 2, 15, and 19, and further teaches
receiving a query where the query is applicable to the first data item and to the instance of the first data item (in response to a query sent by a user, i.e. receiving a query, the system retrieves splits from the vector database which are similar to the query, i.e. where the query is applicable to the first data item and to the instance of the first data item [0156],[0170],[0181],[0183]);
(claim 20) to forward the query to the dynamically prioritized similarity search component (the system uses context-optimized retrieval techniques, such as a search query, to search a large corpus of documents and retrieve a subset of document chunks related to the user query, such as through a vector similarity search, i.e. to forward the query to a…similarity search component, using vector databases where the document chunks are stored, and where the SCORE-RAG system scores and ranks the chunks based on relevance to a particular query, i.e. dynamically prioritized similarity search component [0152],[0156-9],[0170], [0172],[0181],[0183],[0189]);
querying the first and second vectorized databases for the first and second embeddings (the system uses context-optimized retrieval techniques, such as a search query, to retrieve a subset of documents related to the user query, i.e. querying…for the first and second embeddings, using vector databases where the document chunks are stored, i.e. querying the first and second vectorized databases [0152],[0156-9],[0170],[0172],[0181],[0183][0189]);
returning the first weighting assigned to the first data item (the SCORE-RAG system employs a relevance scorer to assess the importance of each chunk, i.e. the first weighting assigned to the first data item, and the chunks are scored and ranked in response to the user query, i.e. returning the first weighting [0170],[0172],[0181]);
returning the second weighting assigned to the instance of the first data item (the SCORE-RAG system employs a relevance scorer to assess the importance of each chunk, i.e. the second weighting assigned to the instance of the first data item, and the chunks are scored and ranked in response to the user query, i.e. returning the second weighting [0170],[0172],[0181]); and
returning one of the first data item or the instance of the first data item associated with the higher weighting (the highly scored chunks, such as the top chunk, may be sent to the LLM for processing to create results, i.e. returning one of the first data item or the instance of the first data item associated with the higher weighting [0157],[0172],[0181]).
(claim 20) to append 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 (an augmentation engine may be operable to combine retrieved information with the original query, i.e. append the query with…information, such as the vector representation of the highest scoring chunks, i.e. associated with the one of the first data item or the instance of the first data item associated with the higher weighting, and relevant metadata and contextual information, i.e. append the query with augmented context information [0170],[0172],[0181-2]); and
(claim 20) to passing query with the augmented context information to the LLM (the retrieved splits and relevant associated metadata and context information are sent to the LLM along with the query, where the query is augmented [0156],[0181-3]).
Regarding claims 5 and 17, Madisetti teaches claims 4 and 16, and further teaches
appending the query with augmented context information associated with one of the first data item or the instance of the first data item associated with the higher weighting (an augmentation engine may be operable to combine retrieved information with the original query, i.e. appending the query with…information, such as the vector representation of the highest scoring chunks, i.e. associated with one of the first data item or the instance of the first data item associated with the higher weighting, and relevant metadata and contextual information, i.e. appending the query with augmented context information [0170],[0172],[0181-2]);
passing the augmented context information to the LLM (the retrieved splits and relevant associated metadata and context information are sent to the LLM along with the query, where the query is augmented [0156],[0181-3]);
querying the LLM with the appended query (the LLM generates an answer based on the query and context information [0156],[0181-3]); and
returning a response from the LLM based on the appended query (the LLM generates an answer based on the query and context information, which is returned [0153],[0156],[0162],[0164],[0181-3]).
Regarding claim 8, Madisetti teaches claim 1, and further teaches
assigning a first weighting to the first data item to be added to the LLM includes assigning the first weighting to the first data item based on a time of generation of the first data item (a time-weighted scorer is operable to incorporate timestamps into chunk scoring, i.e. assigning the first weighting to the first data item, where the timestamp includes the creation date of the chunk, i.e. based on a time of generation of the first data item, and more recent chunks may be ranked higher [0181],[0195]); 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 time of generation of the instance of the first data item (a time-weighted scorer is operable to incorporate timestamps into chunk scoring, i.e. assigning the second weighting to the instance of the first data item, where the timestamp includes the creation date of the chunk, i.e. based on a time of generation of the instance of the first data item, and more recent chunks may be ranked higher [0181],[0195]).
Regarding claim 9, Madisetti teaches claim 8, and 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 most recent time of generation (a time-weighted scorer is operable to incorporate timestamps into chunk scoring, i.e. first or second weighting, where the timestamp includes the creation date of the chunk, i.e. time of generation, and more recent chunks may be ranked higher, i.e. determining which…is a higher weighting includes determining which of the first or second weightings is based on a most recent time of generation [0181],[0195]).
Regarding claim 10, Madisetti teaches claim 1, and further teaches
assigning a first weighting to the first data item to be added to the LLM includes assigning the first weighting to the first data item based on a first origin of information describing the first data item (the citation analyzer identifies and evaluates citations within the received document that contribute to the overall relevance scoring of different text segments, i.e. a first origin of information describing the first data item, where the citation analysis score is determined by the most influential or frequently cited sources, i.e. assigning the first weighting to the first data item based on a first origin of information [0181],[0183]); 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 (the citation analyzer identifies and evaluates citations within the received document that contribute to the overall relevance scoring of different text segments, i.e. a second origin of information describing the instance of the first data item, where the citation analysis score is determined by the most influential or frequently cited sources, i.e. assigning the second weighting to the instance of the first data item based on a second origin of information [0181],[0183]).
Regarding claim 11, Madisetti teaches claim 10, and 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 (the citation analyzer identifies and evaluates citations within the received document that contribute to the overall relevance scoring of different text segments, i.e. origin of information, where the citation analysis score is determined by the most influential or frequently cited sources, i.e. based on a higher priority origin of information, and the document having the greatest citation analysis score is selected for further processing, i.e. determining which of the first or second weightings is based on a higher priority origin of information [0181],[0183]).
Claim Rejections - 35 USC § 103
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) 6, 12, and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Madisetti, in view of Cameron et al. (U.S. PG Pub No. 2025/0036777), hereinafter Cameron.
Regarding claim 6, Madisetti teaches claim 4, and further teaches
determining which of the first or second weightings is a higher weighting based on a most recent time of generation (a time-weighted scorer is operable to incorporate the timestamps, such as creation date, last modified date, or publication date, into chunk scoring, where more recent chunks may be ranked higher [0181]);
determining which of the first or second weightings is a higher weighting is based on a higher priority origin of information (SCORE-RAG may perform citation analysis, which identifies the most influential or frequently cited sources, i.e. based on a higher priority origin of information, where the chunks of document having the greatest citation analysis score are selected for further processing, i.e. determining which of the first or second weightings is a higher weighting [0183]).
While Madisetti provides adjusting a document chunk score based on different contextual information, Madisetti does not specifically teach determining weighting based on a higher vulnerability risk, and thus does not teach
determining which of the first or second weightings is a higher weighting based on a higher vulnerability risk.
Cameron, however, teaches determining which of the first or second weightings is a higher weighting 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. based on a higher vulnerability risk, where the weight is generated for input information, i.e. determining which of the first or second weightings is a higher weighting, and a RAG system can provide supporting information to an LLM for generate an output, such as an exploit or patch for a vulnerability [0055-6],[0069-71]).
Madisetti and Cameron are analogous art because they are from a similar field of endeavor in weighting information in a RAG system for LLM output generation. 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 adjusting a document chunk score based on different contextual information teachings of Madisetti with generating weights based on security vulnerability of a specific platform as taught by Cameron. It would have been obvious to combine the references to improve the performance of an LLM by generating longer, more complex, more nuanced, and/or more specific outputs, such as providing responses similar to what an attacker would do (Cameron [0072]).
Regarding claim 12, Madisetti teaches claim 1.
While Madisetti provides adjusting a document chunk score based on different contextual information, Madisetti does not specifically teach determining weighting based on a higher vulnerability risk, and thus does not teach
assigning a first weighting to the first data item to be added to the LLM includes assigning the first weighting to the first data item 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 assigning a first weighting to the first data item to be added to the LLM includes assigning the first weighting to the first data item 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. assigning the first weighting to the first data item 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, and a RAG system can provide supporting information to an LLM for generate an output, such as an exploit or patch for a vulnerability [0055-6],[0069-71]); 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, and a RAG system can provide supporting information to an LLM for generate an output, such as an exploit or patch for a vulnerability [0055-6],[0069-71]).
Madisetti and Cameron are analogous art because they are from a similar field of endeavor in weighting information in a RAG system for LLM output generation. 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 adjusting a document chunk score based on different contextual information teachings of Madisetti with generating weights based on security vulnerability of a specific platform as taught by Cameron. It would have been obvious to combine the references to improve the performance of an LLM by generating longer, more complex, more nuanced, and/or more specific outputs, such as providing responses similar to what an attacker would do (Cameron [0072]).
Regarding claim 13, Madisetti in view of Cameron teaches claim 12, and Cameron 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 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 a 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, where the weight is generated for input information, and a RAG system can provide supporting information to an LLM for generate an output, such as an exploit or patch for a vulnerability [0055-6],[0069-71],[0099]).
Where the motivation to combine is the same as previously presented.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Madisetti, in view of Rajagopal (U.S. PG Pub No. 2012/0005218), hereinafter Rajagopal.
Regarding claim 7, Madisetti teaches claim 1, and 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 (the highly scored chunks, such as the top chunk, may be sent to the LLM for processing to create results, i.e. updating the LLM with one of the first data item or the instance of the first data item associated with the higher weighting [0157],[0172],[0181]), processing the first and second weightings according to at least one of:
maintaining the first and second weightings until a condition is met (scoring and ranking of the chunks are performed in real-time in response to a particular user query, i.e. maintaining the first and second weightings until a condition is met [0170-2]).
While Madisetti provides calculating the score and ranking in response to a particular user query, Madisetti does not specifically teach resetting the weightings after a determined period of time or converging weightings, and thus does not teach
resetting the first and second weightings after a determined period of time;
converging the first and second weightings into a single weighting.
Rajagopal, however, teaches resetting the first and second weightings after a determined period of time (the score for an article in a knowledge base of articles, i.e. first and second weightings, can be re-determined at a specific time, such as every 30 days or every 7 days, i.e. resetting…after a determined period of time [0045],[0056],[0061]);
converging the first and second weightings into a single weighting (for two articles “A” and “B”, each article has its own score, i.e. first and second weightings, and if article “A” is referenced in article “B”, the referring article’s score can be considered as a parameter for determining the score of article “A”, i.e. converging…into a signal weighting [0045],[0047],[0051-2]).
Madisetti and Rajagopal are analogous art because they are from a similar field of endeavor in scoring and ranking information for search retrieval. 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 calculating the score and ranking in response to a particular user query teachings of Madisetti with re-determining scores at a specific time and using scores of other articles as part of the score calculation for a particular article as taught by Rajagopal. It would have been obvious to combine the references to improve the performance of search engines using search analytics (Rajagopal [0041]).
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