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 .
The following action is in response to the original filing of 06/29/2024.
Claims 1-20 are pending and have been considered below.
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.
Claim 4 is 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.
Regarding claim 4, claim 4 recites the limitations: “wherein an LLM from the plurality of LLMs” and “scanning content data items in an instructive cluster”. Regarding the plurality of LLM language, parent claim 1 states that there are “one or more LLMs” and not explicitly a plurality of LLMs; there exists an interpretation wherein only one LLM exists and not a plurality as required by the language of claim 4 which renders claim 4 unclear. Regarding the content data items in an instructive cluster, there is a lack of antecedent basis for “content data items in an instructive cluster”; parent claim 1 generates “a plurality of instructive clusters associated with the data instance” and makes no mentions of content data items in an instructive cluster.
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 abstract ideas without significantly more.
Regarding claims 1, 8 and 15:
Step 1, MPEP 2106.03:
These limitations have been determined, under Step 1, to be statutory categories of invention:
A computerized system comprising: one or more computer processors; and computer memory storing computer-useable instructions that, when used by the one or more computer processors, cause the one or more computer processors to perform operations [..] (claim 1)
A method, the method comprising: [..] (claim 8)
One or more computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising: (claim 15)
Step 2A Prong One MPEP 2106.04, 2106.04(a):
These limitations represent, under Step 2A Prong One, mental processes such as concepts that can be practically performed in the human mind, or by a human using pen and paper as a physical aid, including observations, evaluations, judgments and opinions, MPEP 2106.04(a)(2)(III), for example a user can find the most common features of a dataset, cluster data of the dataset based on the common features and present topic for each cluster including a category and summary of each cluster and then present results:
[..] generating a plurality of data item embeddings for the data items in the data instance; [..] (claim 1)
[..] reducing dimensionality of the plurality of data item embeddings; [..] (claim 1)
[..] using one or more .. clustering techniques and plurality of data item embeddings having reduced dimensions, generating a plurality of instructive clusters associated with the data instance; [..] (claims 1, 8 and 15)
[..] using a topic modeling technique, generating a topic annotation for the plurality of instructive clusters; [..] (claims 1 and 15)
[..] using .. the plurality of instructive clusters with topic annotations, generating a plurality of annotated clusters, wherein an annotated cluster comprises a category annotation and a summary annotation; [..] (claims 1 and 15)
[..] and using the plurality of annotated clusters, generating a data analysis pipeline output comprising a data instance assessment [..] (claims 1 and 15)
Similarly, these limitations represent, under Step 2A Prong One, mental processes such as concepts that can be practically performed in the human mind, or by a human using pen and paper as a physical aid, including observations, evaluations, judgments and opinions, MPEP 2106.04(a)(2)(III), for example a user can find the most common features of a dataset, cluster data of the dataset based on the common features, filter the cluster by a category to generate a knowledge graph and then present results:
[..] generating a filtered plurality of instructive clusters based on filtering the plurality of instructive clusters using a filtering criteria comprising a representative data feature parameter; [..] (claims 8 and 15)
[..] using .. the filtered plurality of instructive clusters, generating a reasoned knowledge graph; [..] (claims 8 and 15)
[..] and using the reasoned knowledge graph associated with the plurality of instructive clusters, generating a data analysis pipeline output comprising a data instance assessment [..] (claims 8 and 15)
Additionally, these limitations represent, under Step 2A Prong One, mental processes such as concepts that can be practically performed in the human mind, or by a human using pen and paper as a physical aid, including observations, evaluations, judgments and opinions, MPEP 2106.04(a)(2)(III), for example a user present plural results together:
[..] and using the first analysis pipeline output and the second data analysis pipeline output, generating a merged data analysis pipeline output [..] (claim 15)
Step 2A Prong Two, MPEP 2106.04(d):
These limitations represent, under Step 2A Prong Two, mere instructions to implement the abstract idea using generic computing tools, MPEP 2106.05(f):
[..] A computerized system comprising: one or more computer processors; and computer memory storing computer-useable instructions that, when used by the one or more computer processors, cause the one or more computer processors to perform operations: [..] (claim 1)
[..] One or more computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising: [..] (claim 15)
[..] using one or more unsupervised clustering techniques [..] (claims 1, 8 and 15)
[..] using one or more large language models (LLMs) and the plurality of instructive clusters with topic annotations [..] (claims 1 and 15)
[..] using a plurality of artificial intelligence (AI) agents and the filtered plurality of instructive clusters, [..] (claims 8 and 15)
These limitations represent, under Step 2A Prong Two, mere data gathering, MPEP 2106.05:
[..] accessing a data instance comprising data items, wherein the data instance is associated with a dataset; [..] (claims 1, 8 and 15)
Step 2B, MPEP 2106.05:
These limitations are considered, under Step 2B, insignificant extra-solution activity as being recited at a high level of generality, MPEP 2106.05(d):
[..] A computerized system comprising: one or more computer processors; and computer memory storing computer-useable instructions that, when used by the one or more computer processors, cause the one or more computer processors to perform operations: [..] (claim 1)
[..] One or more computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising: [..] (claim 15)
[..] using one or more unsupervised clustering techniques [..] (claims 1, 8 and 15)
[..] using one or more large language models (LLMs) and the plurality of instructive clusters with topic annotations [..] (claims 1 and 15)
[..] using a plurality of artificial intelligence (AI) agents and the filtered plurality of instructive clusters, [..] (claims 8 and 15)
These limitations are considered, under Step 2B, insignificant extra-solution activity of data gathering/selecting a particular type of data, MPEP 2106.05(g):
[..] accessing a data instance comprising data items, wherein the data instance is associated with a dataset; [..] (claims 1, 8 and 15)
Regarding dependent claims 2, 9 and 16, these dependent claim further recites limitations that the unsupervised clustering technique includes HDBSCAN and k-means clustering executed using UMAP (claims 2, 9 and 16). The analysis incorporates the Step analysis of its respective parent. These limitations represent, under Step 2A Prong One, mathematical concepts such as mathematical relationships, mathematical formulas or equations, or mathematical calculations, MPEP 2106.04(a)(2)(I). Under Steps 2A Prong Two and Step 2B, all respective limitations are part of the abstract idea.
Regarding dependent claims 3 and 17, these dependent claim further recites limitations that the modeling technique is based on TF-IDF using seed words additionally requiring that the topic annotation is a top word summarizing subject matter (claims 3 and 17). The analysis incorporates the Step analysis of its respective parent. These limitations represent, under Step 2A Prong One, mathematical concepts such as mathematical relationships, mathematical formulas or equations, or mathematical calculations, MPEP 2106.04(a)(2)(I). The additional limitations represent, under Step 2A Prong Two, mere instructions to apply at a high level of generality (MPEP 2106.05); under Step 2B, these additional limitations are considered mere instructions to apply to obtain a solution/outcome (MPEP 2106.05(f)).
Regarding dependent claims 5, 7, 12, 14 and 20, these dependent claim further recites limitations that include filtering the clusters based on a feature, generating a knowledge graph and presenting output, additionally requiring utilizing a plurality of AI agents including Reflexion and RJMC-LLM agents (claim 5); data pipeline output is associated with a pipeline based on clustering and topic modeling that enable generating of reasoned knowledge graphs based on graph reasoning and inference deduction loops, additionally requiring the pipeline being unsupervised and using LLMs (claims 7, 14 and 20); and generating topic modeling annotations to generate annotated clusters including category and summary annotations and output a result, additionally requiring the use of LLMs (claim 12). The analysis incorporates the Step analysis of its respective parent. These limitations represent, these limitations represent, under Step 2A Prong One, mental processes such as concepts that can be practically performed in the human mind, or by a human using pen and paper as a physical aid, including observations, evaluations, judgments and opinions, MPEP 2106.04(a)(2)(III). The additional limitations represent, under Step 2A Prong Two, mere instructions to implement the abstract idea using generic computing tools, MPEP 2106.05(f); under Step 2B, insignificant extra-solution activity as being recited at a high level of generality, MPEP 2106.05(d).
Regarding dependent claims 4, 6, 11, 13 and 18, these dependent claim further recites limitations that involve additional manipulation of the input data items or output data pipelines (claims 4 and 6) using particular iterative techniques (claims 11, 13 and 18). The analysis incorporates the Step analysis of its respective parent. Additionally, these limitations represent, these limitations represent, under Step 2A Prong One, mental processes such as concepts that can be practically performed in the human mind, or by a human using pen and paper as a physical aid, including observations, evaluations, judgments and opinions, MPEP 2106.04(a)(2)(III). Under Steps 2A Prong Two and Step 2B, all respective limitations are part of the abstract idea.
Regarding dependent claims 10 and 19, these dependent claim additionally recites that the AI agents are based on Reflexion agents or RJMC-LLM agents. The analysis incorporates the Step analysis of its respective parent. The additional limitations represent, under Step 2A Prong Two, mere instructions to implement the abstract idea using generic computing tools, MPEP 2106.05(f); under Step 2B, insignificant extra-solution activity as being recited at a high level of generality, MPEP 2106.05(d).
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 3, 4, 7, 8, 11-15, 17, 18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Howard et al., US 2025/0053617 A1 effective filing 08/09/2023 [HOWARD], in view of Mouyra, US 2023/0351257 A1 published 11/02/2023 [MOURYA].
Regarding claim 1, HOWARD discloses a computerized system comprising:
one or more computer processors; and
computer memory storing computer-useable instructions that, when used by the one or more computer processors, cause the one or more computer processors to perform operations, the operations comprising:
accessing a data instance comprising data items, wherein the data instance is associated with a dataset (pp. 59: transaction records data items from a plurality);
generating a plurality of data item embeddings for the data items in the data instance (pp. 47: transaction string transformed into vector, pp. 60: transaction strings are vectorized);
reducing dimensionality of the plurality of data item embeddings (pp. 44: dimensionality reduction algorithm, pp. 61: group vectors into groups within vector space);
using one or more unsupervised clustering techniques and plurality of data item embeddings having reduced dimensions, generating a plurality of instructive clusters associated with the data instance (pp. 44: clustering algorithms, pp. 60, pp. 61: generate clusters from groupings representing “islands of meaning”);
using a topic modeling technique, generating a topic annotation for the plurality of instructive clusters (pp. 44: topic modeling algorithm, pp. 84: label categorized/group transactions exemplars); and
using the generated topic annotation and the plurality of instructive clusters, generating a data analysis pipeline output comprising a data instance assessment (pp. 18: using processing methodology comprising steps from transaction, algorithm to output, i.e. pipeline output, to determine patters of correlation, Fig. 7A).
While HOWARD teaches the use of one or more LLMs (pp. 66: ex. NLP ML model) to generate at least category annotations and summary annotations (pp. 52: model generates mappings include category and/or enhances transaction strings, broadly a summary), Howard fails to explicitly disclose wherein an LLM and the plurality of instructive clusters with topic annotations are used to generate annotated clusters comprising the category and summary annotations.
MOURYA discloses a data processing pipeline method involving preprocessing stages such as clustering (pp. 16), an analogous art. In particular, MOURYA discloses using one or more LLM agent and a plurality of topic annotated clusters to generate category and summary annotated clusters (pp. 184-190: utilizing NLP models to mine topic modeled text for text categorization and/or text summarization and presenting as output, pp. 16: ex. generating label of intent categories).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD and MOURYA before them before the effective filing of the claimed invention to combine the generating of annotated clusters comprising category and summary annotations using at least an LLM and clusters of data items, as taught by MOURYA, with the use of one or more LLMs to generate category and summary annotations of HOWARD, predictably yielding the result of using one or more LLMs and the plurality of instructive clusters with topic annotations of the combination of HOWARD and MOURYA to generate a plurality of annotated clusters comprising category and summary annotations. One would have been motivated to make this combination in order to identify previously hidden or underlying concepts or intents in the data that were not previously explicitly labeled for use in helping the user discover new suggestions and recommendations as the user navigates information, as suggested by MOURYA (pp. 184-190, pp. 228, pp. 268-269, Fig. 9).
Regarding claim 3, HOWARD and MOURYA disclose the system of claim 1, and MOURYA further discloses wherein the topic modeling technique is based on class-based TF-IDF (Term Frequency-Inverse Document Frequency) matrix and seed words that are representative of a data features to be extracted from the plurality of instructive clusters (pp. 208: extracting important keywords using TF-IDF algorithm), wherein a topic annotation is a top word assigned to an instructive cluster, the top word summarizing subject matter encapsulated within its contents (pp. 209: topic modeling generates dominant topic to use as label).
Regarding claim 4, HOWARD and MOURYA disclose the system of claim 1, and MOURYA further discloses wherein an LLM from the plurality of LLMs supports scanning content data items in an instructive cluster, wherein the data items are scored and prioritized based on predefined categories associated with scanned content of the data items (pp. 230: using predefined categories to score/prioritize data items, Fig. 9).
Regarding claim 7, HOWARD and MOURYA disclose the system of claim 1, and MOURYA further discloses wherein the data analysis pipeline output is associated with a data analysis pipeline, wherein the data analysis pipeline is an unsupervised learning pipeline based on clustering, topic modeling, and large language models that enable generation of reasoned knowledge graphs based on a plurality of graph-based reasoning and inference agents that execute induction and deduction loops (pp. 100, pp. 184-190, pp. 212, pp. 225, pp. 231-233, Fig. 4: data analysis pipelines including unsupervised learning based on clustering, topic modeling, and LLMs to generate graphs based on inference agents executing induction/deduction loops).
Regarding claim 8, HOWARD discloses a method, the method comprising:
accessing a data instance comprising data items, wherein the data instance is associated with a dataset (pp. 59: transaction records data items from a plurality);
using one or more unsupervised clustering techniques, generating a plurality of instructive clusters associated with the data instance (pp. 44: clustering algorithms, pp. 60, pp. 61: generate clusters from groupings representing “islands of meaning”);
generating a filtered plurality of instructive clusters based on filtering the plurality of instructive clusters using a filtering criteria comprising a representative data feature parameter (pp. 44: filtering, pp. 56: filtering using information);
using the filtered plurality of instructive clusters, generating a reasoned knowledge output (pp. 44: generate decision tree); and
using the reasoned knowledge output associated with the plurality of instructive clusters, generating a data analysis pipeline output comprising a data instance assessment (pp. 18: pipeline processing methodology using algorithms to produce output, Fig. 7A).
HOWARD fails to disclose wherein the reasoned knowledge output is a graph generated using a plurality of AI agents.
MOURYA discloses a data processing pipeline method involving preprocessing stages such as clustering (pp. 16), an analogous art. In particular, MOURYA discloses using artificial intelligence agents and a group of filtered clusters to generate reasoned knowledge graphs to create pipeline outputs (pp. 232: generate reasoned knowledge graph structure, pp. 252: tree and/or graph, pp. 100, pp. 212, pp. 225: trained agent to classify queries and provide relevant responses, pp. 228-233, Fig. 9).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD and MOURYA before them before the effective filing of the claimed invention to combine the generating of reasoned knowledge graphs using AI agents, as taught by MOURYA, with the generation of the reasoned knowledge output of HOWARD, predictably yielding the result of generating a reasoned knowledge graph of HOWARD and MOURYA using AI agents. One would have been motivated to make this combination in order to identify previously hidden or underlying concepts or intents in the data that were not previously explicitly labeled for use in helping the user discover new suggestions and recommendations as the user navigates information, as suggested by MOURYA (pp. 184-190, pp. 228, pp. 268-269, Fig. 9).
Regarding claim 11, HOWARD and MOURYA disclose the method of claim 8, and MOURYA further discloses wherein generating the reasoned knowledge graph is based on an induction-deduction technique associated with executing iterative loops and refinements for continuous learning and validation (pp. 100, pp. 184-190, pp. 212, pp. 225, pp. 231-233, Fig. 4: data analysis pipelines generate graphs based on inference agents iteratively executing induction/deduction loops).
Regarding claim 12, HOWARD and MOURYA disclose the method of claim 8, and MOURYA further discloses the method further comprising:
using a topic modeling technique, generating a topic annotation the plurality of instructive clusters (pp. 184: topic modeling techniques to discover latent topics);
using one or more Large Language Models (LLMs) and the plurality of instructive clusters with topic annotations, generating a plurality of annotated clusters, wherein an annotated cluster comprises a category annotation and a summary annotation (pp. 184-190: utilizing NLP models to mine topic modeled text for text categorization and/or text summarization and presenting as output, pp. 16: ex. generating label of intent categories); and
generating a second data analysis pipeline output comprising a second data instance assessment (Fig. 9).
Regarding claim 13, HOWARD and MOURYA disclose the method of claim 8, and MOURYA further discloses applying a graph-cut community detection algorithm to partition the reasoned knowledge graph into clusters by iteratively splitting or merging nodes to identify densely connected communities in the reasoned knowledge graph (pp. 252: clustering the graph nodes into hierarchical relationships representing a taxonomy, yielding parent-child node relationships, pp. 267: user explores and discovers new relationships which yield new intent categories yielding new ways to explore, i.e. iterative).
Regarding claim 14, HOWARD and MOURYA disclose the method of claim 8, and MOURYA further discloses wherein the data analysis pipeline output is associated with a data analysis pipeline, wherein the data analysis pipeline is an unsupervised learning pipeline based on clustering, topic modeling, and large language models that enable generation of reasoned knowledge graphs based on a plurality of graph-based reasoning and inference agents that execute induction and deduction loops (pp. 100, pp. 184-190, pp. 212, pp. 225, pp. 231-233, Fig. 4: data analysis pipelines including unsupervised learning based on clustering, topic modeling, and LLMs to generate graphs based on inference agents executing induction/deduction loops).
Regarding claim 15, HOWARD discloses one or more computer-storage media having computer-executable instructions embodied thereon that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising:
accessing a data instance comprising data items, wherein the data instance is associated with a dataset (pp. 59: transaction records data items from a plurality);
using one or more unsupervised clustering techniques, generating a plurality of instructive clusters associated with the data instance (pp. 44: clustering algorithms, pp. 60, pp. 61: generate clusters from groupings representing “islands of meaning”);
using a topic modeling technique, generating a topic annotation of the plurality of instructive clusters (pp. 44: topic modeling algorithm, pp. 84: label categorized/group transactions exemplars);
using one or more Large Language Models (LLMs) and the plurality of instructive clusters with topic annotations, generating a plurality of annotated clusters, wherein an annotated cluster comprises a category annotation and a summary annotation;
using the generated topic annotation and the plurality of instructive clusters, generating a first data analysis pipeline output comprising a first data instance assessment (pp. 18: using processing methodology comprising steps from transaction, algorithm to output, i.e. pipeline output, to determine patters of correlation, Fig. 7A);
generating a filtered plurality of instructive clusters based on filtering the plurality of instructive clusters using a filtering criteria comprising a representative data feature parameter (pp. 44: filtering, pp. 56: filtering using information);
using the filtered plurality of instructive clusters, generating a reasoned knowledge output (pp. 44: generate decision tree);
using the reasoned knowledge output associated with the plurality of instructive clusters, generating a second data analysis pipeline output comprising a second data instance assessment (pp. 18: pipeline processing methodology using algorithms to iteratively produce output, Fig. 7A); and
using the first data analysis pipeline output and the second data analysis pipeline output, generating a merged data analysis pipeline output (pp. 48: combining plural model pipelines to provide a combined prediction output).
While HOWARD teaches the use of one or more LLMs (pp. 66: ex. NLP ML model) to generate at least category annotations and summary annotations (pp. 52: model generates mappings include category and/or enhances transaction strings, broadly a summary), Howard fails to explicitly disclose wherein an LLM and the plurality of instructive clusters with topic annotations are used to generate annotated clusters comprising the category and summary annotations and fails to explicitly disclose wherein the reasoned knowledge output is a graph generated using a plurality of AI agents.
MOURYA discloses a data processing pipeline method involving preprocessing stages such as clustering (pp. 16), an analogous art. In particular, MOURYA discloses using one or more LLM agent and a plurality of topic annotated clusters to generate category and summary annotated clusters (pp. 184-190: utilizing NLP models to mine topic modeled text for text categorization and/or text summarization and presenting as output, pp. 16: ex. generating label of intent categories) and using artificial intelligence agents and a group of filtered clusters to generate reasoned knowledge graphs to create pipeline outputs (pp. 232: generate reasoned knowledge graph structure, pp. 252: tree and/or graph, pp. 100, pp. 212, pp. 225: trained agent to classify queries and provide relevant responses, pp. 228-233, Fig. 9).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD and MOURYA before them before the effective filing of the claimed invention to combine the generating of annotated clusters comprising category and summary annotations using at least an LLM and clusters of data items and generating of reasoned knowledge graphs using AI agents, as taught by MOURYA, with the use of one or more LLMs to generate category and summary annotations and knowledge reason outputs of HOWARD, predictably yielding the result of using one or more LLMs and the plurality of instructive clusters with topic annotations of the combination of HOWARD and MOURYA to generate a plurality of annotated clusters comprising category and summary annotations and generating a reasoned knowledge graph of HOWARD and MOURYA using AI agents. One would have been motivated to make this combination in order to identify previously hidden or underlying concepts or intents in the data that were not previously explicitly labeled for use in helping the user discover new suggestions and recommendations as the user navigates information, as suggested by MOURYA (pp. 184-190, pp. 228, pp. 268-269, Fig. 9).
Regarding claim 17, HOWARD and MOURYA disclose the media of claim 15, and MOURYA further discloses wherein the topic modeling technique is based on class-based TF-IDF (Term Frequency-Inverse Document Frequency) matrix and seed words that are representative of a data features to be extracted from the plurality of instructive clusters (pp. 208: extracting important keywords using TF-IDF algorithm), wherein a topic annotation is a top word assigned to an instructive cluster, the top word summarizing subject matter encapsulated within its contents (pp. 209: topic modeling generates dominant topic to use as label).
Regarding claim 18, HOWARD and MOURYA disclose the media of claim 15, and MOURYA further discloses wherein generating the reasoned knowledge graph is based on an induction-deduction technique associated with executing iterative loops and refinements for continuous learning and validation (pp. 100, pp. 184-190, pp. 212, pp. 225, pp. 231-233, Fig. 4: data analysis pipelines generate graphs based on inference agents iteratively executing induction/deduction loops)
Regarding claim 20, HOWARD and MOURYA disclose the media of claim 15, and MOURYA further discloses wherein the merged data analysis pipeline output is associated with a data analysis pipeline, wherein the data analysis pipeline is an unsupervised learning pipeline based on clustering, topic modeling, and LLMs that enable generation of reasoned knowledge graphs based on a plurality of graph-based reasoning and inference agents that execute induction and deduction loops (pp. 100, pp. 184-190, pp. 212, pp. 225, pp. 231-233, Fig. 4: data analysis pipelines including unsupervised learning based on clustering, topic modeling, and LLMs to generate graphs based on inference agents executing induction/deduction loops).
Claims 2, 9 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over HOWARD in view of MOURYA and in further view of Jha, US 2024/0346086 effective filing of 04/13/2023 [JHA].
Regarding claim 2, HOWARD and MOURYA disclose the system of claim 1, and MOURYA further discloses wherein the one or more unsupervised clustering techniques include Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) clustering and Spherical k-means clustering (pp. 24), the one or more unsupervised clustering techniques are executed on reduced data item embeddings associated with dimensionality reduction (pp. 187).
However, neither HOWARD nor MOURYA explicitly disclose wherein the dimensionality reduction algorithm uses UMAP.
JHA discloses methods for categorizing data items using clustering (pp. 10-11), an analogous art. In particular, JHA discloses executing unsupervised clustering techniques such as HDBSCAN on reduced data associated with dimensionality using UMAP (pp. 52: utilize uniform manifold approximation and projection as a substitution for PCA, pp. 85).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD, MOURYA and JHA before them before the effective filing to combine the use of UMAP to reduce dimensionality for executing clustering techniques, as taught by JHA, when executing the clustering techniques on a reduced dimensionality set of HOWARD and MOURYA. One would have been motivated to make this combination in order to minimize information loss due to compression, as suggested by JHA (pp. 51-52).
Regarding claims 9 and 16, claims 9 and 16 each recite limitations similar to claim 2 and are similarly rejected.
Claims 5, 6, 10 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over HOWARD in view of MOURYA and in further view of Huang, Yuxuan. "Evaluating and enhancing large language models for conversational reasoning on knowledge graphs." arXiv preprint arXiv:2312.11282 (2023) [HUANG].
Regarding claim 5, HOWARD and MOURYA disclose the system of claim 1, and MOURYA further discloses the operations further comprising:
generating a filtered plurality of instructive clusters based on filtering the plurality of instructive clusters using a filtering criteria comprising a representative data feature parameter (pp. 228: search filter to narrow intent categories);
using artificial intelligence (AI) agents and the filtered plurality of instructive clusters, generating a reasoned knowledge graph (pp. 232: generate reasoned knowledge graph structure, pp. 252: tree and/or graph, pp. 100, pp. 212, pp. 225: trained agent to classify queries and provide relevant responses), and
using the reasoned knowledge graph, generating a second data analysis pipeline output comprising a second data instance assessment (pp. 228-233: present iterative data analysis outputs based on filtering functions and/or graph structure, Fig. 9, pp. 248-252, Fig. 10).
Neither HOWARD nor MOURYA disclose wherein the AI agents include one or more of the following: a Reflexion agent and Reversible Jump Markov Chain-LLM agent.
HUANG discloses methods for processing information into a knowledge graph using agents (page 2 pp. 4: “We introduce LLM-ARK, an effective framework that employs LLM as an Agent for Reasoning on Knowledge Graphs. We employ LLMs as agent and express the Large model KG inference task as a reinforcement learning sequential decision-making problem, and using a Full-Textual-Environment prompt to aggregate multiscale inputs.”), an analgous art. In particular, HUANG discloses that it was well known to use agents, such as Reflexion agents, when applying LLM agents to a task (page 3 pp. 2: “Shinn et al. [2023] introduce a novel framework, Reflexion, that strengthens linguistic agents through linguistic feedback, rather than updating weights. The Reflexion agent verbally reflects on task feedback signals, and then stores its reflection text in an episodic memory buffer to make better decisions in subsequent trials.”), and that determining the knowledge graph reasoning using a Markov process applied over variable states, i.e. reversible jump Markov chain (page 4 pp. 6: “KG reasoning naturally reduces to a finite horizon, deterministic partially observed Markov decision process that lie on a KG G. We formulate KG reasoning as a Markov Decision Process (MDP) described by a five-tuple (S,O,A,T,R,γ): • State. S is an infinite set of environment states, which encode information stored in Working Memory, including task background tb, user query q, dialog history h, current entity vc, path history ph, current step t, The normal state is represented using a six-tuple: S = (tb,q,h,vc,ph,t).”).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD, MOURYA and HUANG before them before the effective filing of the claimed invention to combine the use of a reflexion agent or reverse jump Markov chain when determining the knowledge graph, as suggested by HUANG, when generating/determining the knowledge graph of HOWARD and MOURYA, predictably yielding the result of the AI agents of HOWARD, MOURYA and HUANG applying at least one of a reflexion or reversible jump Markov chain. One would have been motivated to make this combination in order to provide a more efficiently traversable knowledge graphs achieved using a a more effective framework, as suggested by HUANG (page 4 pp. 5: “To overcome this challenge, our study employs a policy gradient model that efficiently traverses the KG to select relationships and ultimately achieves the target, demonstrating proficiency in multi-hop reasoning.”, page 2 pp. 4: “We introduce LLM-ARK, an effective framework that employs LLM as an Agent for Reasoning on Knowledge Graphs. We employ LLMs as agent and express the Large model KG inference task as a reinforcement learning sequential decision-making problem, and using a Full-Textual-Environment prompt to aggregate multiscale inputs.”).
Regarding claim 6, HOWARD, MOURYA and HUANG disclose the system of claim 5, and HOWARD further discloses the operations further comprising, using the first data analysis pipeline output and the second data analysis pipeline output, generating a merged data analysis pipeline output (pp. 48: combining plural model pipelines to provide a combined prediction output).
Regarding claim 10, HOWARD and MOURYA disclose the method of claim 8.
HOWARD and MOURYA fail to disclose wherein the AI agents including one or more of the following: a Reflexion agent and Reversible Jump Markov Chain-LLM agent.
HUANG discloses methods for processing information into a knowledge graph using agents (page 2 pp. 4: “We introduce LLM-ARK, an effective framework that employs LLM as an Agent for Reasoning on Knowledge Graphs. We employ LLMs as agent and express the Large model KG inference task as a reinforcement learning sequential decision-making problem, and using a Full-Textual-Environment prompt to aggregate multiscale inputs.”), an analgous art. In particular, HUANG discloses that it was well known to use agents, such as Reflexion agents, when applying LLM agents to a task (page 3 pp. 2: “Shinn et al. [2023] introduce a novel framework, Reflexion, that strengthens linguistic agents through linguistic feedback, rather than updating weights. The Reflexion agent verbally reflects on task feedback signals, and then stores its reflection text in an episodic memory buffer to make better decisions in subsequent trials.”), and that determining the knowledge graph reasoning using a Markov process applied over variable states, i.e. reversible jump Markov chain (page 4 pp. 6: “KG reasoning naturally reduces to a finite horizon, deterministic partially observed Markov decision process that lie on a KG G. We formulate KG reasoning as a Markov Decision Process (MDP) described by a five-tuple (S,O,A,T,R,γ): • State. S is an infinite set of environment states, which encode information stored in Working Memory, including task background tb, user query q, dialog history h, current entity vc, path history ph, current step t, The normal state is represented using a six-tuple: S = (tb,q,h,vc,ph,t).”).
Therefore it would have been obvious to one having ordinary skill in the art and the teachings of HOWARD, MOURYA and HUANG before them before the effective filing of the claimed invention to combine the use of a reflexion agent or reverse jump Markov chain when determining the knowledge graph, as suggested by HUANG, when generating/determining the knowledge graph of HOWARD and MOURYA, predictably yielding the result of the AI agents of HOWARD, MOURYA and HUANG applying at least one of a reflexion or reversible jump Markov chain. One would have been motivated to make this combination in order to provide a more efficiently traversable knowledge graphs achieved using a a more effective framework, as suggested by HUANG (page 4 pp. 5: “To overcome this challenge, our study employs a policy gradient model that efficiently traverses the KG to select relationships and ultimately achieves the target, demonstrating proficiency in multi-hop reasoning.”, page 2 pp. 4: “We introduce LLM-ARK, an effective framework that employs LLM as an Agent for Reasoning on Knowledge Graphs. We employ LLMs as agent and express the Large model KG inference task as a reinforcement learning sequential decision-making problem, and using a Full-Textual-Environment prompt to aggregate multiscale inputs.”).
Regarding claim 19, claim 19 recites limitations similar to claim 10 and is similarly rejected.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Commons; Michael Lamport
US 11514305 B1
INTELLIGENT CONTROL WITH HIERARCHICAL STACKED NEURAL NETWORKS
Gou; Liang et al.
US 12118439 B2
MIXED-INITIATIVE MACHINE LEARNING SYSTEMS AND METHODS FOR DETERMINING SEGMENTATIONS
D'Agostino; Dino Paul
US 12682179 B2
RESPONSE DETERMINATION BASED ON CONTEXTUAL ATTRIBUTES AND PREVIOUS CONVERSATION CONTENT
Miao; Yi et al.
US 20180322122 A1
RECOMMENDATIONS FOR ONLINE SYSTEM GROUPS
Parlak; Ayse et al.
US 20200159925 A1
AUTOMATED MALWARE ANALYSIS THAT AUTOMATICALLY CLUSTERS SANDBOX REPORTS OF SIMILAR MALWARE SAMPLES
Kesavan; Krishnakumar Gopalan et al.
US 20240354169 A1
SYSTEM AND METHOD FOR DYNAMIC RESOURCE MANAGEMENT AND ALLOCATION FOR CLUSTER NETWORKS
Churgin; Matthew A. et al.
US 20240371367 A1
AUTOMATED CALL SUMMARIZATION BASED ON FILTERED UTTERANCES
Seth; Yashu et al.
US 20240403694 A1
GROUP RECOMMENDATION BASED ON TOPIC CLUSTERING
Yao; Weiran et al.
US 20250045567 A1
SYSTEMS AND METHODS FOR LANGUAGE AGENT OPTIMIZATION
Salman; Tamer
US 20250061195 A1
ARTIFICIAL INTELLIGENCE SECURITY ENGINE IN A SECURITY MANAGEMENT SYSTEM
Wan; Mengting et al.
US 20250094538 A1
DATASET CLUSTERING VIA LANGUAGE MODEL PROMPTS
Pickman; Stephen et al.
US 20250117486 A1
CLUSTERING OF HIGH DIMENSIONAL DATA AND USE THEREOF IN CYBER SECURITY
Talwalkar; Priya Saurabh et al.
US 20250173607 A1
ACTIONABLE INSIGHT GENERATION FOR INCIDENTS GENERATED FROM EVENTS
Yamaura; Sahim et al.
US 20250307663 A1
METHOD AND SYSTEM FOR GENERATING KNOWLEDGE GRAPH
Liu; Zhiwei et al.
US 20250348704 A1
SYSTEMS AND METHODS FOR CONTROLLABLE ARTIFICIAL INTELLIGENCE AGENTS
Chen, Hai-bo
CN 111694958 A
MICRO-BLOG TOPIC CLUSTERING METHOD BASED ON WORD VECTOR AND SINGLE-PASS FUSION
Bao, Junwei, et al. "Constraint-based question answering with knowledge graph." Proceedings of COLING 2016, the 26th international conference on computational linguistics: technical papers. 2016.
Abels, Patrick, et al. "Focusing knowledge-based graph argument mining via topic modeling." arXiv preprint arXiv:2102.02086 (2021).
Clay, Robert, et al. "Real-time agent-based crowd simulation with the reversible jump unscented Kalman filter." Simulation Modelling Practice and Theory 113 (2021): 102386.
Stanik, Christoph, Tim Pietz, and Walid Maalej. "Unsupervised topic discovery in user comments." 2021 IEEE 29th International Requirements Engineering Conference (RE). IEEE, 2021.
Hidayat, Alam Ahmad, et al. "BERT-based topic modeling approach for malaria research publication." 2022 International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS). IEEE, 2022.
Welivita, Anuradha, and Pearl Pu. "Heal: A knowledge graph for distress management conversations." Proceedings of the AAAI conference on artificial intelligence. Vol. 36. No. 10. 2022.
Weng, Min-Hsien, Shaoqun Wu, and Mark Dyer. "Identification and visualization of key topics in scientific publications with transformer-based language models and document clustering methods." Applied Sciences 12.21 (2022): 11220.
De Raedt, Maarten, et al. "IDAS: Intent discovery with abstractive summarization." Proceedings of the 5th Workshop on NLP for Conversational AI (NLP4ConvAI 2023). 2023.
Ropiak, Norbert, et al. "How to select samples for active learning? Document clustering with active learning methodology." 2023 27th International Conference on Engineering of Complex Computer Systems (ICECCS). IEEE, 2023.
Syed, Shahbaz, et al. "Indicative summarization of long discussions." Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.
Zhang, Yuwei, Zihan Wang, and Jingbo Shang. "Clusterllm: Large language models as a guide for text clustering." Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.
Cui, Jiaxi, et al. "Chatlaw: A multi-agent collaborative legal assistant with knowledge graph enhanced mixture-of-experts large language model." arXiv preprint arXiv:2306.16092 (2023).
Harikrishnan, Karunya, M. Malathi, and K. B. Sundharakumar. "Topic-driven contractual language understanding and summarization: An integrated approach for simplifying legal documents." 2024 4th International Conference on Intelligent Technologies (CONIT). IEEE, 2024.
Tarekegn, Adane Nega. "Large language model enhanced clustering for news event detection." arXiv preprint arXiv:2406.10552 (2024).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANDREW L TANK whose telephone number is (571)270-1692. The examiner can normally be reached Monday-Thursday 9a-6p.
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, Matthew Ell can be reached at 571-270-3264. 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.
/ANDREW L TANK/ Primary Examiner, Art Unit 2141
