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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 5/22/2026 has been entered.
Status
This action is in response to the amendment filed on 5/22/2026. Claims 1, 2, 4-7, 10-12, 14-17, 25 are pending. Claims 1, 11, are amended. Claims 21-25 have been added. No claims are currently cancelled.
Response to Arguments
Applicant’s arguments, see pages 9-10, with respect to have been fully considered and are persuasive. The previous 112(a) rejections have been withdrawn.
Applicant's arguments filed 5/22/2026 have been fully considered but they are not persuasive. The applicant has argued the claims in view of Streamlined Example 2 “The present claims are analogous in the same legally relevant manner. Like the robotic arm example, the claims here do not merely recite an abstract objective such as analyzing information or generating questions. Instead, they recite a specific technical system involving a pre-trained language model, a knowledge graph constructed from document content, and a defined processing pipeline in which the knowledge graph is used to guide chunking and preserve contextual relationships. The claims further specify how this preserved contextual information is used by the PLM to generate predicted questions. Thus, the claims define a concrete mechanism by which a computer system processes data, rather than merely invoking abstract reasoning. Moreover, just as the robotic arm example improves the operation of a physical system (by achieving smoother motion through a defined control mechanism), the present claims improve the operation of a computer system-specifically, the functioning of a language model when processing large documents. The technical problem addressed (loss of contextual information due to input size constraints) and the technical solution (knowledge graph-guided chunking with preservation of contextual patterns) parallel the structure of the eligible example, where a specific mechanism is used to achieve improved system performance. Accordingly, under the same reasoning applied by the USPTO in Example 2, the present claims are not directed to an abstract idea and are patent eligible.” The examiner respectfully disagrees. The example was found eligible not because it recited a “specific technical system” processing data through a pipeline. The claim was found eligible because it was directed to a particular machine (a robotic arm, physical end effector, sensor, and control system). The mathematical relationship was not the improvement to the invention but merely used in the operation of the physical structure. Applicant’s claims do not recite an analogous physical structure. Claim 1 is a data processing method, claim 11 only recites generic one or more hardware processors. The claims do not recite the use of a sensor, physical actuator, or a control system adjusting a physical process. Constructing a knowledge graph, chunking text, encoding via a graph attention network, and generating output text are all data transformations performed on generic computing hardware. Further, the applicant does not disclose nor claim how the knowledge graph informs chuck boundaries.
The claimed improvement is to the abstract idea and not to the functionality of the computer. The applicant appears to be arguing that the invention is improving the functioning of language models when processing large documents. However, applicant’s specification and claims describe the improvement as more accurate and complete predicted question generation, not a change in how the underlying computer or model operates (see abstract). The courts have distinguished between claims that improve a computer’s technical operation and claims that use conventional computing components to make an abstract analytical process more accurate. Preserving question patterns so that a pretrained language model produces better-tailored questions is similar to Electric Power Group which is an improvement in the quality of the analytical output, not in the computer that performs the analysis.
Claim 1 recites that the knowledge graph is “used to guide chunking of the document such that question patterns reflected within the document are preserved across resulting chunks, thereby avoiding loss of information included in the document during chunking, wherein, as a result of chunking the document using the knowledge graph.” This claim limitation merely recites that patterns are preserved, not the specific rule, algorithm, or decision criteria by which the knowledge graph determines chunk boundaries. No specificity appears in claim 1 for how the knowledge graph informs chunk boundaries, the claim states only the desired outcome.
The applicant has argued that “Claims 21 and 22 recite detailed operations for constructing the knowledge graph using knowledge triples and transforming those triples into a structured graph representation including nodes and edges, and further modifying that representation by converting edges into nodes and introducing multiple edge types (default, reverse, and self-edge). These limitations define a particular graph construction and transformation process that operates on structured data representations and cannot be performed in the human mind. Rather than merely organizing information, these claims specify how relationships within document data are encoded into a machine-readable graph structure, thereby reflecting a concrete data-processing technique and not a mental process or method of organizing human activity.” The examiner respectfully disagrees. The relevant inquiry is on the underlying steps extracting a subject, relation and object from text, creating two nodes, creating an edge, connecting the nodes with the edge are steps a person could perform using pen and paper. Extracting a subject relation object from a sentence is human reading comprehension. Representing a relationship as two labeled dogs connected by a line is also something a human can do. Claiming nodes and edges is simply data structure for an act that a person could do with pen and paper. Claim 22 is a way of relabeling and redrawing a relationship diagram so that a relation is itself treated as a node with new connecting lines of designated type. No new information is derived or computation performed. The applicant does not identify any claimed rule, threshold, or algorithm for how the data is extracted from unstructured text, nor any specific technical benefit from the labeling. Claims 21-22 only recite a generic sequence of triple extraction and node/edge labeling without any claimed technical mechanism producing an analogous technical benefit. Applicant’s asserts without additional analysis that these operations cannot practically be performed in the human mind. This is a bare assertion without any reference to claim language. Based on the broadest reasonable interpretations of the claims a human is able to read a sentence, identify the subject, and sketch a relationship between two points.
The applicant has argued “Claim 23 recites generating and providing a prompt template that includes document context and a question output placeholder to the PLM. This limitation introduces a specific mechanism for interfacing with the language model, defining how input is structured and provided to the model for downstream processing. This is a concrete implementation detail of how the PLM operates on input data, not an abstract concept, and further ties the claimed invention to a particular technological environment involving controlled prompt construction for machine learning systems.” The examiner respectfully disagrees. Limiting an abstract idea to a particular field of use or technological environment does not integrate the idea into a practical application. The claim does not recite any particular structure, format, syntax, or algorithm for the template, only that is includes context from the document and a question output placeholder. This appears to be similar to form letter. The claim recites no details it merely recites the tasking without the functional means. The applicant does not identify anything about the recited template that departs from its conventional usage, nor any specific technical benefit from the particular template structure claimed. Unlike claims found eligible for reciting a specific rule based process, claim 23 does not specify any rule by which document context is selected, truncated, or formatted for insertion into a template, nor does it site any technical details about placement or context of the question output placeholder. Therefore the claim does not integrate the underling abstract idea into a practical application.”
The applicant has argued “Claims 24 and 25 further recite extracting embeddings generated from the PLM, computing a similarity metric relative to a learned history of embeddings, and assigning generated questions to a cluster plane of analysts based on that similarity and a historical behavior profile including sentiment and question-type features. These limitations introduce additional machine learning operations involving vector embeddings, similarity computations, and clustering based on learned models. Such operations require computational processing of high-dimensional data representations and cannot be practically performed mentally. Moreover, these steps define how the system evaluates and organizes generated outputs using learned representations, thereby improving the functionality and output relevance of the system.” The examiner respectfully disagrees. Claim 24 recites comparing a new question to a set of prior questions and assessing how alike they are, which appears to be a comparative judgement a person performs whenever they recognize that a new question sounds like one they have heard before. While Claim 25 appears to be categorizing an item into a pre-existing group based on how similar it is to the groups known characteristics. Comparing data, recognizing similarities or patterns, and organizing information into categories is an abstract mental process. The claims do no recite any specific technical architecture that would place the actual claimed operations beyond human mental capability. The claims do not recite any change to how the underlying computer or pretrained language model operates. They recite an additional layer of post processing applied to an already generated question. Assigning a cluster plane to a question based on similarity to historical behavior is a refinement of the abstract idea, not a change to the functionality of the computer.
Applicant’s arguments are not found persuasive. The previous 101 rejection is maintained and updated below.
The applicant has amended the claims to overcome the previous 112, second paragraph rejections. The previous 112, second paragraph rejections have been withdrawn.
The applicant has argued the previous 103 rejections. The applicant has amended the claims which required an updated prior art search. Upon careful review of the prior art, Pappu et al. (US 20180004718 A1) and Jungmeisteris et al. (US 20220374956 A1) do not teach or suggest the invention as a whole, particularly the combination in order of the limitations of identifying a PLM that is tasked with analyzing the document to generate a predicted question of the questioning entity; generating a knowledge graph for the document, wherein the knowledge graph tracks question patterns included into the document; chunking the document using the knowledge graph, wherein the knowledge graph is used to guide chunking of the document such that question patterns reflected within the document are preserved across resulting chunks, thereby avoiding loss of information included in the document during chunking, wherein, as a result of chunking the document using the knowledge graph, the question patterns included within the document are preserved in respective chunks, and wherein the document is chunked based on a model size of the PLM; encoding the knowledge graph…; after the document has been chunked into the respective chunks, tasking the PLM to generate the predicted question of the questioning entity; and causing the PLM to generate the predicted question. While knowledge graphs, document chunking, graph attention networks, and PLM-based text generation are individually known in the art, the prior art of record does not disclose or render obvious the specific These claimed limitations were are not found in the prior art in combination. While knowledge graphs, document chunking, graph attention networks, and PLM based generation are individually known in the art, the prior art of record does not disclose or render obvious the specific ordered combination. The combined teachings of the references do not disclose the required combination of PLM, knowledge graph features for generating the intended outcome, nor is there a reasonable motivation to combine them in a manner that achieves the claimed invention without the benefit of hindsight.
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, 2, 4-7, 10-12, 14-17, 20-25 are rejected under 35 USC 101 because the claimed invention is directed to a judicial exception (i.e. abstract idea) without anything significantly more.
Step 1: Claims 1, 2, 4-7, 10, 21-25, are directed to a method and claims 11- 12, 14-17, 20 are directed to a non-transitory storage medium. Therefore, claims 1, 2, 4-7, 10-12, 14-17, 20-25 are directed to patent eligible categories of invention.
Step 2A Prong 1: The claim(s) recite(s) (mathematical relationships/formulas, mental process or certain methods of organizing human activity). Specifically the independent claims recite:
(a) mental process: as drafted, the claim recites the limitation of preparing a document draft, performing a sentiment analysis on question patterns, performing a sentiment analysis, performing a topic generation process, performing an automated question generation process, another sentiment analysis, and updating a list of topics which is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. That is, other than reciting “by one or more hardware processors,” nothing in the claim precludes the determining step from practically being performed in the human mind. For example, but for the “by one or more hardware processors” language, the claim encompasses the user manually performing a sentiment analysis and updating a list of topics. The mere nominal recitation of a generic processing device does not take the claim limitation out of the mental processes grouping. This limitation is a mental process.
(b) mathematical formula: The claim recites a mathematical concept (which can include a mathematical relationships, mathematical formulas or equations, and mathematical calculations), and in this case a graph attention network encoding step and pretrained language model attention layer which involve underlying mathematic operations. While these are recited functionality the operations they invoke are fundamentally mathematical relationships among data representations. Thus, the claim recites a mathematical concept, mathematical formulas or equations, or mathematical relationships). Note that, in this example, the “encoding” step is determined to recite a mathematical concept because the claim explicitly recites a mathematical formula or calculation.
(c) certain methods of organizing human activity: The claim as a whole recites a method of organizing human activity. The claimed invention is a method that allows for tracking of human question patterns, sentiment, and topics, and structuring predicted future interaction. Allowing users to perform sentiment analysis on a document draft to update a list of topics related which is a method of managing interactions between people. According to the 2019 PEG, “managing personal behavior or relationships or interactions between people” includes social activities, teaching, and following rules or instructions.
With the exception of the “using the at least one processor” language, the claim steps in the context of the claim encompass an abstract idea directed to “Certain Methods of Organizing Human Activity.” Thus, the claim recites an abstract idea.
Dependent claims 2, 4-7, 10, 12, 14-17, 20-22, 25, further narrow the abstract idea identified in the independent claims and do not introduce further additional elements for consideration.
Dependent claims 23-24 will be evaluated under Step 2A, Prong 2 below.
Step 2A, Prong 2: Independent claims 1, 11, do not integrate the judicial exception into a practical application. Claim 1 is a method that recites “a pre-trained language model,” “a knowledge graph”, “chunking,” “a graph attention network.” Claim 11 is a non-transitory storage medium that recites “automated” limitations performed “therein instructions that are executable by one or more hardware processors to perform operations… a pre-trained language model, “a knowledge graph”, “chunking,” “a graph attention network.” The recitation of a processor in claim 11 is merely a generic computer performing generic computing functions, while claim 1 does not recite a machine limitation. These additional elements are mere instructions to implement an abstract idea using a computer in its ordinary capacity, or merely uses the computer as a tool to perform the identified abstract idea. Use of a computer or other machinery in its ordinary capacity for performing the steps of the abstract idea or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., certain methods of organizing human activity) does not integrate a judicial exception into a practical application. See MPEP 2106.05(f).
Dependent Claims 23 and 24 include the additional element of the “PLM.” Use of a computer or other machinery in its ordinary capacity for performing the steps of the abstract idea or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., certain methods of organizing human activity) does not integrate a judicial exception into a practical application. See MPEP 2106.05(f).
Therefore, the additional elements of the independent claims, when considered both individually and in combination, are not sufficient to prove integration into a practical application.
Dependent claims 2, 4-7, 10, 12, 14-17, 20-22, 25, further narrow the abstract idea identified in the independent claims and do not introduce further additional elements for consideration, which does not integrate the judicial exception into a practical application.
Therefore, the additional elements of the dependent claims, when considered both individually and in the context of the independent claims, are not sufficient to prove integration into a practical application.
Step 2B: Independent claims 1, 11 do not comprise anything significantly more than the judicial exception. As can be seen above with respect to Step 2A, Prong 2, Claim 1 is a method at most comprising various steps that are “automated.” Claim 1 is a method that recites “a pre-trained language model,” “a knowledge graph”, “chunking,” “a graph attention network.” Claim 11 is a non-transitory storage medium that recites “automated” limitations performed “therein instructions that are executable by one or more hardware processors to perform operations… a pre-trained language model, “a knowledge graph”, “chunking,” “a graph attention network.” These additional elements are mere instructions to implement an abstract idea using a computer in its ordinary capacity, or merely uses the computer as a tool to perform the identified abstract idea. Use of a computer or other machinery in its ordinary capacity for performing the steps of the abstract idea or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., certain methods of organizing human activity) is not anything significantly more than the judicial exception. See MPEP 2106.05(f).
Dependent Claims 23 and 24 include the additional element of the “PLM.” Use of a computer or other machinery in its ordinary capacity for performing the steps of the abstract idea or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., certain methods of organizing human activity) is not anything significantly more than the judicial exception. See MPEP 2106.05(f).
The additional elements of the independent claims, when considered both individually and in combination, do not comprise anything significantly more than the judicial exception.
Dependent claims 2, 4-7, 10, 12, 14-17, 20-22, 25, further narrow the abstract idea identified in the independent claims and do not introduce further additional elements for consideration, which is not anything significantly more than the judicial exception.
The additional elements of the dependent claims, when considered both individually and in the context of the independent claims, are not anything significantly more than the judicial exception.
Accordingly, claims 1, 2, 4-7, 10-12, 14-17, 20-25, are rejected under 35 USC 101.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1, 2, 4-7, 10-12, 14-17, 20-25 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention. The applicant has amended claims 1 and 11 to include the language of “chunking the document using the knowledge graph, wherein the knowledge graph is used to guide chunking of the document such that question patterns reflected within the document are preserved across resulting chunks, thereby avoiding loss of information included in the document during chunking, wherein, as a result of chunking the document using the knowledge graph, the question patterns included within the document are preserved in respective chunks, and wherein the document is chunked based on a model size of the PLM.” Although the applicant has support in paragraph 68 of the originally filed disclosure for wherein the document is chunked based on the size of the PLM. However, the applicant does not have support in the originally filed disclosure for the knowledge graph controlling how chunk boundaries are drawn. Appropriate correction is required.
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 1, 2, 4-7, 10-12, 14-17, 20-25 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 1 and 11 recite the limitation “the document.” Although there is support for “a document draft.” There is insufficient antecedent basis for this limitation in the claim.
Claims 1 and 11 recite the limitations “a predicted question” and “a set of predicted questions.” It is unclear how generating one “predicted question” related to producing the previously recited “set of predicted questions.” Is the set the aggregate of multiple singular predicted question generations? Clarification is requested.
Claim 21 recites “constructing the knowledge graph.” Claim 1 recited “generating a knowledge graph for a document.” Claim 1 does not use the word constructing so it is unclear if claim 21 is further limiting the generating step of claim 1, or introducing an additional separate constructing step. Clarification is requested.
Claim 24 recites “a generated question.” Claim 1 recites “the predicted question.” It is unclear whether a generated question in claim 24 is intended to refer back to the predicted question of 1 or whether it is meant to introduce a new separate question distinct from claim 1’s predicted question. Clarification is requested.
Claim 25 recites “the generated question” and “the computed similarity metric.” There is insufficient antecedent basis for these limitations in the claim.
The dependent claims inherit the rejection of the claim from which they depend.
Other pertinent prior art includes Pappu et al. (US 20180004718 A1) discloses providing facts related to user generated content, including comments posted on webpages. Jungmeisteris et al. (US 20220374956 A1) which discloses a dynamic, accurate, and timely customer sentiment analysis .Prieto (US 20140164036 A1) which discloses a program manager can monitor or track sentiment data related to a capital program. Blanchflower et al. (US 20160071119 A1) which discloses generating a proposed sentiment result associated with a document, the proposed sentiment result being generated based on a rule set applied to the document. Zhu et al. ( US 20220358594 A1) which discloses a machine learning model that can be trained to predict one or more financial indicators using earnings call transcripts augmented with counterfactual information.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JAMIE H AUSTIN whose telephone number is (571)272-7363. The examiner can normally be reached Monday, Tuesday, Thursday, Friday 7am-2pm.
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, Brian Epstein can be reached at (571) 270 5389. 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.
JAMIE H. AUSTIN
Examiner
Art Unit 3625
/JAMIE H AUSTIN/Primary Examiner, Art Unit 3625
Prosecution Insights