DETAILED ACTION
Introduction
This office action is in response to applicant’s amendment filed 2/18/2026. Claims 1-20 are currently pending and have been examined. Applicant’s IDS have been considered. There is no claim to foreign priority.
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 .
Response to Arguments
Applicant’s arguments, see remarks, filed 2/18/2026, with respect to the rejection(s) of claim(s) 1-20 under 35 USC 103 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of the previously cited prior art and further in view of Madhavan et al. (see rejections below).
Portions of applicant's arguments filed 2/18/2026 have been fully considered but they are not persuasive. More specifically, applicant argues, with respect to claims 1, 11 and 18,
“Baldua does not teach or suggest "generating ... one or more prompt elements by associating context data with the one or more knowledge graph data objects." Instead, the prompt in the Baldua reference is described as user input that "cause[s] the large language model 116 to generate and output a query execution plan" for executing a "query [to] one or more of the data resources 134," which include knowledge graphs." Id. at paragraphs [0053], [0054], [0061], and [0064]. In other words, Baldua describes using a plan generation prompt to query knowledge graphs - not a prompt element generated by associating context data with the one or more knowledge graph data objects. Accordingly, Baldua does not teach or suggest "generating, ... based on a query input, one or more prompt elements by associating context data with the one or more knowledge graph data objects as recited in claims 1, 11, and 18, either previously or amended herewith. Accordingly, Applicant respectfully requests reconsideration and withdrawal of the rejection for at least this reason alone.”
However, the Examiner does not concur with the applicant’s arguments, where Baldua explicitly teaches, generating a prompt, as this is not argued. Baldua further teaches wherein generating the prompt comprises context data, see paragraphs [0047, 0048]-see is prompt generation discussion, which includes generating the prompt based on input and context data. Therefore, these two elements of criteria with respect to the claim are met. Baldua further teaches in paragraphs [0048, 0080], as previously cited, in order to generate the prompt, the final prompt or query uses the input classification associated with the knowledge graph, as the data resources, thus as described and see in the illustration of Fig. 1, the input and context data as associated with the data resources, which contains the knowledge graph, and thus there is a clear associated between these elements, and the subsequently generated prompt is based on each of these respective elements. The Examiner notes the applicant’s argument that Baldua teaches outputting a query execution plan for executing a query to one or more of the data resources, 134, however, the applicant lacks taking not of the entire prompt generation element, which involves the teaching above, and the communication between elements in Fig. 1A item 110 and 134. Wherein the input, context and knowledge graph data as associated are all used in generating a prompt, which can then be used to query a knowledge data source in order to generate a response/answer. Therefore, the corresponding arguments are deemed non-persuasive, wherein the Examiner notes the applicant has appeared to dismiss the input, context and knowledge graph which are used to generate a prompt (zero or few shot prompts) which are then subsequently used as a query to data resources.
The Examiner further notes, Baldua with Krishnan with Madhavan (as newly cited), explicitly detail the amended limitations in their entirety as combined in the teachings as seen below. The current rejection based on the applicant’s amendments.
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) 1, 6-8, 10, 11, 15, 16 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Baldua et al. (Baldua, US 2025/0110957) in view of Krishnan et al. (Krishnan, US 2025/0005050) and further in view of Madhavan et al. (Madhavan, US 2025/0245226).
As per claim 1, Baldua teaches a computer-implemented method comprising:
receiving, by one or more processors, one or more knowledge graph data objects comprising (i) a plurality of nodes associated with a plurality of topics, a plurality of entities, or a plurality of documents and (ii) a plurality of edges between the plurality of nodes (paragraphs [0072, 0064, 147]-his knowledge graph, Fig. 6 item 600, 634, 632, his topics, entities and concepts as part of his knowledge graph, see his edges between nodes, Fig. 8, his processing device, with respect to processor, hereinafter);
generating, by the one or more processors and based on a query input, one or more prompt elements by associating context data with the one or more knowledge graph data objects (paragraph [0025]-his prompt), wherein the context data is further associated with one or more query topics, one or more query entities, or one or more query documents (ibid, see also, paragraph [0025, 0033, 0034, 0047, 0048, 0064, 0080, 0143-0159]-his user history, data sources, associated context data, including topics, or query entities or documents);
providing, by the one or more processors, a prompt comprising the one or more prompt elements to a natural language processing machine learning model [to receive one or more subgraph data objects from the natural language processing machine learning model] (paragraphs [0027-0029]-his GAI is designed using a LLM, neural network and deep learning, for performing NLP tasks, and his explicit providing of a prompt to this GAI model in order to generate a response/answer, this answer/response is coming from the LLM, GAI configured to generate a response, based on training and knowledge graph data); and
providing, by the one or more processors, one or more answer outputs based on the one or more subgraph data [objects] (ibid-his plan generation prompt, sent to the LLM, based on the subgraph data, Fig. 4-his prompt execution, and abstract, paragraphs [0004, 0027, 0042, 0047, 0061-0073, 0101-0111]-all detailing the response/answer outputs based on the subgraph data, Figs. 3-6-the response stemming from the subgraph dependency graph, which processes only graph elements relevant to subset of functions).
Baldua does not explicitly refer or teach subgraph data objects, as taught by Krishnan, his subgraph data objects (paragraph [0213, 0212], Figs. 2B and 6, and hereinafter the subgraph data as the subgraph data objects, as combined below).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Baldua and Krishnan to combine the prior art element of providing answer outputs based on context data associated with knowledge graph data objects as taught by Baldua with the subgraph data objects as taught by Krishnan as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be providing a result, response and answer based on traversing subgraph data objects, (ibid, Krishnan, see also abstract).
The above combination lacks explicitly teaching that which Madhavan teaches, providing, by the one or more processors, a prompt comprising the one or more prompt elements to a natural language processing machine learning model to receive one or more subgraph data objects from the natural language processing machine learning model (paragraphs [0029, 0051-0064, 0151]-his knowledge graph construction system, including a natural language processing machine learning model, within his graph component, which generates the graph elements comprising subgraph data objects, and corresponding prompt, based on an input and contextual data, provided to the artificial intelligence machine learning system to receive from the natural language processing machine learning model subgraph data objects, wherein the machine learning models produce domain specific graphs, and return graph components as accessed, and used to generate an answer/output).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Baldua and Krishnan and Madhavan to combine the prior art element of providing answer outputs based on context data associated with knowledge graph data objects as taught by Baldua with the subgraph data objects as taught by Krishnan with receiving one or more subgraph data objects from a natural language processing machine learning model as taught by Madhavanas each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be providing a result, response and answer based on traversing subgraph data objects, wherein a natural language machine learning model is used to generate and construct a knowledge graph, therefore allowing the system to receive subgraph objects when prompted (ibid, Krishnan, see also abstract, ibid-Madhavan-see his nodes and edges, used to answer the prompts discussion, the nodes and edges coming from, generated by, the natural language processing machine learning model).
As per claim 6, Baldua further makes obvious the computer-implemented method of claim 1, wherein the natural language processing machine learning model comprises a transformer machine learning model (ibid, his transformers, Chat-GPT and BERT used in his NLP models, paragraph [0067]).
As per claims 7 and 15, Baldua further makes obvious the computer-implemented method of claim 1, wherein the one or more prompt elements further comprise (i) one or more historic queries, or (ii) one or more prompt templates (ibid-see claim 1, prompt discussion, see also paragraphs [0025, 0156, 0157]-see his query history utilized in prompt construction, and also his prompt templates).
As per claims 8 and 16, Baldua further makes obvious the computer-implemented method of claim 7, wherein generating the one or more [subgraph] data objects comprises:
generating the prompt based on one of the one or more prompt templates (ibid, paragraph [0101, 0099-0104]-his subgraph and corresponding data components, and generation prompt, prompt templates);
generating a question based on the one or more prompt templates and one or more of (a) the context data, (b) the one or more knowledge graph data objects, or (c) the one or more historic queries (ibid, his prompt and query generated based on the prompt templates, knowledge graph data, historic queries, and further input into LLM, Figs. 1A-Fig. 3); and
generating, using the natural language processing machine learning model, an answer based on the question (ibid-see claim 1, answer/response discussion, his transformer, GPT/BERT, NLP model, LLM and generated response based on the above, context data and generated query/prompt Fig. 7A items 702-718).
Baldua lacks explicitly teaching that which Madhavan teaches, generating the one or more subgraph data object (see claim 1, generating his knowledge graph discussion, his generated prompt, and knowledge graph data objects, and generated answer using the natural language processing machine learning model.
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Baldua and Krishnan and Madhavan to combine the prior art element of providing answer outputs based on context data associated with knowledge graph data objects as taught by Baldua with the subgraph data objects as taught by Krishnan with generating the one or more subgraph data objects, comprising generating the prompt, a question and answer, as described above, using the natural language processing machine learning model as taught by Madhavanas each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be providing a result, response and answer based on traversing subgraph data objects, wherein a natural language machine learning model is used to generate and construct a knowledge graph, therefore allowing the system to receive subgraph objects when prompted (ibid, Krishnan, see also abstract, ibid-Madhavan-see his nodes and edges, used to answer the prompts discussion, the nodes and edges coming from, generated by, the natural language processing machine learning model).
As per claim 10, Baldua further makes obvious the computer-implemented method of claim 1, wherein generating the one or more answer outputs comprises:
traversing the one or more subgraph data objects (ibid-see claim 1, subgraph discussion, Baldua, paragraph [0143, 0144]-his traversed graph structure discussion, see also Khrishnan, paragraphs [0212, 0213]-traversal of sub-graph data objects); and
identifying one or more entities or one or more documents from the one or more subgraph data objects that are relevant to the query input based on the traversal (ibid-the traversal to identify entities, via queries, thus relevant to queries, based on the traversal, Krishnan, as similarly motivated and combined above).
As per claim 11, claim 1 sets forth limitations similar to claim 1 and is thus rejected under similar reasons and rationale, wherein the system is deemed to embody the method, such that Baldua with Krishnan with Madhavan make obvious a system comprising one or more processors and
one or more non-transitory computer readable media storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations (paragraphs [0213-0219]-see his system, computer readable media, processors, memory discussion):
receiving one or more knowledge graph data objects comprising (i) a plurality of nodes associated with a plurality of topics, a plurality of entities, or a plurality of documents and (ii) a plurality of edges between the plurality of nodes (ibid-see claim 1, corresponding and similar limitation); generating, based on a query input, one or more prompt elements by associating context data with the one or more knowledge graph data objects, wherein the context data is further associated with one or more query topics, one or more query entities, or one or more query documents (ibid); providing a prompt comprising the one or more prompt elements to a natural language processing machine learning model to receive one or more subgraph data objects from the natural language processing machine learning model (ibid); and providing one or more answer outputs based on the one or more subgraph data objects (ibid).
As per claim 18, claim 18 sets forth limitations similar to claim 1 and is thus rejected under similar reasons and rationale, wherein the non-transitory computer-readable storage media is deemed to embody the method, such that Baldua with Krishnan with Madhavan make obvious one or more non-transitory computer-readable storage media storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising (paragraphs [0213-0219]-see his system, processors, memory, and machine readable medium discussion): receiving one or more knowledge graph data objects comprising (i) a plurality of nodes associated with a plurality of topics, a plurality of entities, or a plurality of documents and (ii) a plurality of edges between the plurality of nodes (ibid-see claim 1, corresponding and similar limitation); generating, based on a query input, one or more prompt elements by associating context data with the one or more knowledge graph data objects, wherein the context data is further associated with one or more query topics, one or more query entities, or one or more query documents (ibid); providing a prompt comprising the one or more prompt elements to a natural language processing machine learning model to receive one or more subgraph data objects from the natural language processing machine learning model (ibid); and providing one or more answer outputs based on the one or more subgraph data objects (ibid).
Claim(s) 2-4, 12-14 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Baldua et al. (Baldua, US 2025/0110957) in view of Krishnan in view of Madhavan, as applied to claim 1 above, and further in view of Wang et al. (Wang, NERank+: a graph-based approach for entity ranking in document collections).
As per claims 2, 12 and 19, Baldua further makes obvious the computer-implemented method of claim 1, wherein: (i) the one or more knowledge graph data objects are previously generated based on one or more document-topic-entity relationship features that are associated with the plurality of topics, the plurality of entities, and the plurality of documents (ibid, paragraph [0147, 0064, 0072]-his knowledge graphs as previously generated and stored, and queried, with corresponding topics, entities, relationships between them and weights, similarity measurements and other scores to describe the relationships),
but lack explicitly teaching, that which Wang teaches, and (ii) the one or more document-topic-entity relationship features are generated by generating one or more of (a) a document-topic distribution matrix (see Wang, pages 507-512, section 4…This Fig. 2 Graph, and Entity Aware Topic Modeling discussion, including a Document-topic matrix and weights used to construct the graph, topic-entity common word matrix, topic-document weights, as his relationship/edge weights, entity-topic weights, as the relationship weights between entities and topics, topic random scoring, based on his quality metric, background and topic specificity scoring), (b) a topic- entity-common word matrix (ibid), (c) a plurality of topic-document weights (ibid), (d) a plurality of entity- topic weights (ibid), (e) a plurality of topic randomness scores (ibid), (f) a plurality of entity quality scores (ibid), or (g) a plurality of similarity scores associated with the plurality of entities, the plurality of topics, or the plurality of documents.
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Baldua and Wang to combine the prior art element of providing answer outputs based on context data associated with knowledge graph data objects and subgraph data objects as taught by Baldua with generating knowledge graph data objects as taught by Wang as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be assisting in retrieving ranked data based on a user query and knowledge data object relationships, weights, and objects (Wang, abstract).
As per claims 3 and 13, Baldua with Krishnan with Wang make obvious the computer-implemented method of claim 2, wherein: (i) the one or more knowledge graph data objects are previously generated based on one or more document-topic- entity relationship features that are associated with the plurality of topics, the plurality of entities, and the plurality of documents, and
(ii) the one or more document-topic-entity relationship features are generated by (ibid-see claim 2, corresponding document-topic-entity relationship discussion, Wang, Fig. 2, and matrix discussion):
(a) generating a document-topic distribution matrix that comprises a distribution of the plurality of topics with respect to the plurality of documents (ibid-Wang, previously cited sections and corresponding discussion, his document-topic distribution matrix);
(b) generating a plurality of topic-document weights for the plurality of topics with respect to the plurality of documents based on the document-topic distribution matrix (ibid-see his weights discussion, wrt the corresponding matrix); and
(c) generating a plurality of topic randomness scores for the plurality of topics based on the plurality of topic-document weights (ibid-see his quality metrics, 2 and 3, based on topic specificity, as the topic randomness score, as similarly motivated and combined as seen in claim 2).
As per claims 4 and 14, Baldua with Krishnan with Madhavan with Wang make obvious the computer-implemented method of claim 1, wherein: (i) the one or more knowledge graph data objects are previously generated based on one or more document-topic- entity relationship features that are associated with the plurality of topics, the plurality of entities, and the plurality of documents, and
(ii) the one or more document-topic-entity relationship features are generated by (ibid-see Wang, document topic entity relationship discussion, claims 2, 3):
(a) generating a topic-entity-common word matrix comprising a distribution of the plurality of entities within a plurality of common words that connect the plurality of topics with respect to the plurality of topics in the plurality of documents (ibid-see above topic-common word matrix discussion, Wang sections 3, 4, Fig. 2, Table 1, his collection of common words, and corresponding Topic-entity matrix based on “W”, as his common words),
(b) generating a plurality of entity-topic weights for the plurality of entities with respect to the plurality of topics based on the topic-entity-common word matrix (ibid-see his weights for the edges, R.sub.t.e. based on the topic-entity common word matrix and corresponding weights), and
(c) generating a plurality of entity quality scores for the plurality of entities based the plurality of entity-topic weights (ibid, page 509- section 5, including his quality metrics, and entity richness and quality, based on the entity-topic weights, the motivation to combine, similar and under the same rationale as claim 2).
Claim(s) 9, 17 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Baldua et al. (Baldua, US 2025/0110957) in view of Krishnan, as applied to claim 1 above, and further in view of Man et al. (Man, US 2023/0394605).
As per claims 9, 17 and 20, Baldua teaches the computer-implemented method of claim 1, wherein generating the one or more answer outputs comprises:
determining an amount of the one or more subgraph data objects is zero or exceeds a threshold representative of the query input comprising an ambiguity (paragraph [0032, 0081, 0082, 0188]-his zero results, from the query, as processed, in claim 1, Figs 1A-7A, using his subgraph data);
generating a follow-up question based on the query input, the one or more subgraph data objects, or the one or more knowledge graph data objects (ibid-his modified prompt, based on the initial prompt and query);
receiving a response to the follow-up question (ibid, his user feedback, see also paragraph [0128]-his feedback);
generating one or more follow-up prompt elements based on the response (ibid-his modified prompt as generated); and
[adding] one or more nodes or edges to the one or more knowledge graph data objects based on the one or more follow-up prompt elements (ibid, paragraph [0138, 0190]-his company names, weights, as applied to his knowledge graph, as his data source, for querying the LLM).
Baldua lacks explicitly teaching that which Man teaches, adding one or more nodes or edges to the one or more knowledge graph data objects based on the one or more follow-up prompt elements (paragraph [0080, 0011]-his node or connections added to the knowledge graph, based on prompt elements to the user, to clarify ambiguities).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Baldua, Krishnan and Man to combine the prior art element of providing answer outputs based on context data associated with knowledge graph data objects and subgraph data objects as taught by Baldua with disambiguating ambiguous queries and adding nodes and edges to a knowledge graph as taught by Wang as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be when prompting a user for data, in order to disambiguate and construct a knowledge graph having appropriate nodes and connections for use in answer an ambiguous query (Wang, abstract, paragraph [0011]).
Allowable Subject Matter
Claim 5 is objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
The Examiner notes, the above prior art and cited related prior art, does not explicitly teach or make obvious, “The computer-implemented method of claim 1, wherein the one or more knowledge graph data objects are previously generated based on one or more document-topic- entity relationship features that are associated with the plurality of topics, the plurality of entities, and the plurality of documents, and the one or more document-topic-entity relationship features are generated by generating a plurality of similarity scores for a plurality of mention-context vector and topic-entity-mention vector pairs.”
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure (See PTO-892).
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LAMONT M SPOONER whose telephone number is (571)272-7613. The examiner can normally be reached 8:00 AM -5:00 PM.
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, Daniel Washburn can be reached at (571)272-5551. 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.
/LAMONT M SPOONER/Primary Examiner, Art Unit 2657
5/12/2026
Prosecution Insights