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 .
Status of Claims/Specification
Claims 1-13 were previously pending and subject to a Non-Final Rejection dated May 7, 2026. In the Response, submitted on August 7, 2026, claims 1, and 12-13 were amended, and claim 14 was added; and the specification title was amended. Therefore, claims 1-14 are currently pending and subject to the following final rejection.
Response to Arguments
Applicant’s arguments on Page 11 of the Response, regarding the previous specification objection have been fully considered, and are found persuasive in view of the amended specification title. No new matter was added.
Applicant’s arguments on Pages 11-17 of the Response, regarding the previous rejection of the claims under 35 U.S.C. 101, have been fully considered but are not found persuasive.
On Page 13 of the Response, Applicant states “The ‘generat[ion of] a trained machine learning model for extracting representations’ is not a ‘commercial interactions or fundamental economic practice[ ],’ and does not ‘manag[e] personal behavior or interactions between people.’ Therefore, viewed as a whole, amended independent claim 1 is directed to a specific technique for generating a trained machine-learning model, and not merely a ‘method of organizing human activity’ as defined in M.P.E.P. § 2106.04(a)(2) § II. Therefore, amended claim 1 is patent eligible at Step 2A - Prong 1. Independent claims 12 and 13 have been amended to recite features similar to those discussed above with respect to amended independent claim 1 and are patent eligible at Step 2A - Prong 1 for at least similar reasons. Dependent claims 2-11 are also patent eligible at Step 2A - Prong 1 at least due to their dependence from now patent-eligible claim 1 and further in view of the additional features recited therein.”
Examiner respectfully disagrees and notes that the high-level recitation of “generating a trained machine-learning model for extracting representations” does not preclude the claim from reciting an abstract idea. That is, “generating a trained machine-learning model for extracting representations” recites an additional element (analyzed under Step 2A, Prong Two and Step 2B).
Furthermore, Paras 3-5 of Applicant’s PG Publication explain “Data of product purchases and reviews by users can be handled as bipartite graphs in which the users and the items are represented by nodes, and the purchasing relationship and the review relationship are represented by edges… there is a problem in that a relative relationship indicating which of the user and a persona is likely to purchase an item cannot be learned, and interpretability is poor.” That is, the specification explains that the problem of the invention is in indicating which of the user and a persona is likely to purchase an item. Therefore, as will be further discussed in the detailed rejection below, in light of the specification, the claims recite an abstract idea of a certain method of organizing human activity (e.g., commercial interactions or fundamental economic practices, or managing personal behavior or interactions between people); and a mathematical concept (e.g., mathematical relationship or mathematical calculation).
On Pages 14-16, in discussing Desjardins and Step 2A, Prong Two, Applicant argues “the Specification provides sufficient detail for a person of ordinary skill in the art to recognize the claimed invention as offering an improvement to computer implemented machine [l]earning models. For example, the Specification explains that:
According to the present embodiment described above, additional graph data in which a persona related to a subject node is added as a node is generated, a recommendation loss and a comparison loss are calculated from subject representations and persona representations, and the parameters of the model for extracting representations from graph data are updated based on the recommendation loss and the comparison loss. As a result, a relationship between subjects can be efficiently trained. For example, in training the model for the relative level of a purchase probability between a user and a product, which is a rank, training is performed so that the representations of the user and the representations of a persona are separated, and thus, the model can be trained to obtain a relative relationship indicating which one of the user and the persona is likely to purchase the product. Furthermore, as the representations are separated between the user and personas, it is possible to improve the interpretability of the model as to which persona is likely to purchase the product. That is, it is possible to provide a trained model with an improved interpretability…
Further, the amendments directly tie the recited graph transformation and representation-based processing to the generation of the trained machine-learning model. Thus, unlike mere data gathering, pre-solution activity, or post-solution activity, the recited graph transformation is part of the technological process by which the model is trained and improved. Similar to Ex Parte Desjardins, where the model was improved through iterative adjustment of parameter values and use of the adjusted values in subsequent training iterations, see Desjardins Spec. at [0063] ("The adjusted values of the parameters are then used as current values of the parameters in the next iteration."), amended independent claim 1 recites specific operations that modify and improve the machine-learning model. Therefore, when considered as a whole, amended independent claim 1 recites a specific technical solution involving persona-node generation, representation extraction, and dual-loss training for generating the trained machine learning model for extracting representations. Accordingly, the claim is directed to a technological improvement in machine-learning technology rather than an abstract idea.”
Examiner respectfully disagrees and notes in Ex Parte Desjardins, the specification identified the improvement to machine learning technology by explaining how the machine learning model is trained to learn new tasks while protecting knowledge about previous tasks to overcome the problem of “catastrophic forgetting,” and that the claims reflected the improvement identified in the specification. The specification included disclosures of the effective learning of new tasks in succession in connection with specifically protecting knowledge concerning previously accomplished tasks; allowing the system to reduce use of storage capacity; and the enablement of reduced complexity in the system. Such improvements were tantamount to how the machine learning model itself would function in operation and therefore not subsumed in the identified mathematical calculation.
However, here, Paras. 36-38 of Applicant’s PG Publication merely explain that “the update unit 107 trains the model to be used in the extraction unit 104, depending on the recommendation loss LCF and the comparison loss LPS… Based on the total loss LTOT, the update unit 107 updates the parameters (weight and bias) of the model so that each difference in similarity between the recommendation loss LCF and the comparison loss LPS becomes larger…. the update unit 107 determines whether the updating of the parameters of the model has been completed, which is whether the training of the model has been completed. In the determination as to completion of the parameter updating, it may be determined that the training is completed in a case where the value of a loss becomes equal to or smaller than a threshold…in a case where the parameters have been updated a predetermined number of times… in a case where the absolute value of the update amount of a parameter or the sum of the absolute values becomes a constant value. Note that the determination as to whether the training has been completed is not limited to the above example, and a termination condition normally employed in machine learning may be used. If the training has been completed, the process is terminated. If the training has not been completed, the process returns to step SA1 to continue the processing for the next graph data.”
Examiner respectfully notes that Applicant’s determining whether the training of the model has been completed, by determining whether the updating of the parameters of the model (weight and bias) has been completed (or by a “termination condition normally employed in machine learning may be used”) is not similar to Desjardn’s machine learning model being trained to learn new tasks while protecting knowledge about previous tasks to overcome a technical problem in machine learning of “catastrophic forgetting”, and therefore allowing the system to reduce use of storage capacity, and the enablement of reduced complexity in the system.
Therefore, unlike Desjardin’s improvement to the machine learning technology itself, here Applicant’s claimed high-level recitation of “update the model, using the first…and second loss to generate the trained machine learning model for extracting representations” does merely subsume the identified abstract idea of mathematical calculations (as discussed in the Final Rejection below). That is, “persona-node generation, representation extraction” reflect the abstract idea and the high-level “dual-loss training for generating the trained machine learning model for extracting representations” merely includes instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea. Thus, Applicant’s arguments are not found persuasive.
On Pages 16-17 of the Response, in discussing Step 2B, Applicant further argues “amended claim 1 recites specific technical improvements that go beyond merely implementing an abstract idea on generic computer components. As recognized in the Office Action (Office Action at 7 (indicating that the claims are allowable over the prior art)), the claimed invention provides novel and non-obvious apparatus, methods, and computer-readable media for representation learning and updating of machine learning models. Moreover, amended independent claim 1 recites elements that reflect an inventive concept that is not well-understood, routine, or conventional in the field, as evidenced by the above discussion of the technical advantages provided by the claimed features. Accordingly, the additional elements of amended independent claim 1 amount to significantly more than any alleged abstract idea, because they provide a non-conventional, ordered combination of elements that improves the functioning of a computing system itself, including ‘updating the model, using the first loss and the second loss to generate the trained machine learning model for extracting representations.’ For at least these reasons, Applicant respectfully submits that amended independent claim 1 recites ‘significantly more’ than any alleged judicial exception because the claim demonstrates ‘improvements to any other technology or technical field’ that go ‘beyond generally linking the use of the judicial exception to a particular technological environment.’”
Examiner respectfully disagrees “claim 1 recites specific technical improvements that go beyond merely implementing an abstract idea on generic computer components” or “improves the functioning of a computing system itself” (for the reasons as discussed above). As discussed above generating the trained machine learning model amounts to no more than reciting the words “apply it” (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).
Furthermore, Examiner notes that “the search for an inventive concept should not be confused with a novelty or non-obviousness determination….As made clear by the courts, the “‘novelty’ of any element or steps in a process, or even of the process itself, is of no relevance in determining whether the subject matter of a claim falls within the § 101 categories of possibly patentable subject matter.” …The search for a § 101 inventive concept is thus distinct from demonstrating § 102 novelty…Specifically, lack of novelty under 35 U.S.C. 102 or obviousness under 35 U.S.C. 103 of a claimed invention does not necessarily indicate that additional elements are well-understood, routine, conventional elements. Because they are separate and distinct requirements from eligibility, patentability of the claimed invention under 35 U.S.C. 102 and 103 with respect to the prior art is neither required for, nor a guarantee of, patent eligibility under 35 U.S.C. 101.” Thus, Applicant’s arguments are not found persuasive.
Applicant’s arguments on Page 18 of the Response, regarding the previous rejection of the claims under 35 U.S.C. 112(b), have been fully considered and are found persuasive in view of the amended claims.
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 1-14 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
Claims 1-11 and 14 recite an apparatus (comprising a processor) (i.e., a machine), claim 12 recites a method (i.e., process), and claim 13 recites a non-transitory computer readable medium (i.e., a machine). Therefore, the claims all fall within one of the four statutory categories of invention.
Step 2A, Prong One
Claims 1, 12, and 13 recite generating a model for extracting representations: acquiring graph data in which a plurality of subjects are represented by nodes for each of two or more categories, and a relationship between the subjects is represented by an edge connecting the nodes; generating additional graph data in which one or more personas having representative features of the subjects are added to the graph data as the nodes, wherein generating the additional graph data comprises generating at least one persona node associated with a respective subjective node; extracting subject representations and persona representations from the additional graph data, using a model for extracting representations; calculating a first loss depending on a difference between a similarity between nodes connected by the edge and a similarity between nodes not connected by the edge, using the subject representations and the persona representations, wherein the similarity is determined based on an association between vectors corresponding to the subject representations and the persona representations; calculating a second loss depending on a difference between a similarity between a node of the subject and a node connected to the node of the subject, and a similarity between a node of the persona and a node connected to the node of the persona, using the subject representations and the persona representations, the second loss being configured to distinguish the subject representations from the persona representations; and updating the model, using the first loss and the second loss to generate the model for extracting representations.
The claims as a whole recites a certain method of organizing human activity. The limitations recited above, under broadest reasonable interpretation, recite the abstract idea of a certain method of organizing human activity, e.g., commercial interactions or fundamental economic practices, or managing personal behavior or interactions between people; and mathematical concepts (e.g., mathematical relationships and calculations) . Therefore, the claims recite an abstract idea.
Step 2A, Prong Two
Claims 1, 12 and 13 as a whole amount to no more than reciting the words “apply it” (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).
Claims 1, 12, and 13 recite the additional element of: (i) a trained machine learning model; claim 1 recites the additional element: (ii) an information processing apparatus comprising a processor; and claim 13 recites the additional element: (iii) a non-transitory computer readable medium including computer executable instructions, wherein the instructions, when executed by a processor, cause the processor to perform a method.
The above additional elements of: (i)-(iii) are recited at a high-level of generality such that, when viewed as whole/ordered combination, they amount to no more than reciting the words “apply it” (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).
Accordingly, these additional elements, when viewed as a whole/ordered combination (See Figs. 1 and 8) do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. Thus, claims 1, 12 and 13 are directed to an abstract idea.
Step 2B
As discussed above with respect to Step 2A Prong Two, the additional elements of claims 1, 12 and 13 amount to no more than reciting the words “apply it” (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea. The same analysis applies here in 2B, i.e., reciting the words “apply it” (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea, does not integrate the abstract idea into a practical application at Step 2A or provide an inventive concept at Step 2B.
Therefore, the additional elements discussed above do not integrate the abstract idea into a practical application at Step 2A or provide an inventive concept at Step 2B. Thus, even when viewed as a whole/ordered combination, nothing in the claims add significantly more (i.e., an inventive concept) to the abstract idea. Thus, claims 1, 12 and 13 are also ineligible.
Dependent claims 2-11 and 14 further recite details which merely narrow the previously recited abstract idea limitiaitions. For these reasons, as described above with respect to claim 1, these judicial exceptions are not meaningfully integrated into a practical application or significantly more than the abstract idea. Thus, claims 2-11 and 14 are also ineligible.
Allowable over the Prior Art
Claims 1-14 are allowable over the prior art because the prior art fails to teach or suggest, “calculating a first loss depending on a difference between a similarity between nodes connected by the edge and a similarity between nodes not connected by the edge, using the subject representations and the persona representations, wherein the similarity is determined based on an association between vectors corresponding to the subject representations and the persona representations; calculating a second loss depending on a difference between a similarity between a node of the subject and a node connected to the node of the subject, and a similarity between a node of the persona and a node connected to the node of the persona, using the subject representations and the persona representations, the second loss being configured to distinguish the subject representations from the persona representations; and updating the model, using the first loss and the second loss to generate the trained machine learning model for extraction representations” as recited in the entirety of the independent claims.
The closest prior art for the independent claims includes:
WO2025/170581 to Carbune et al. (hereinafter “Carbune”). Carbune discloses generating a group persona by generating a plurality of persona candidates and inter-persona dynamics for the persona candidates based on a description of desired characteristics for the group persona.
U.S. Patent Application Publication No. 2020/0035002 to Epasto et al. (hereinafter “Epasto”). Epasto discloses receive data describing a first graph, and for each node, of one or more nodes, of the first graph, determine, based at least in part on data describing a second graph, and for each of multiple nodes of the second graph corresponding to the node of the first graph, a representation of a role of the node of the multiple nodes in a community to which the node of the multiple nodes belongs.
CN115470379 to Xu et al. (hereinafter “Xu”). Xu discloses acquiring user original data, and generating a user relation graph according to the user original data. The user relationship graph is constructed by a plurality of user nodes and a plurality of relationship edges, each relationship edge is connected with two user nodes, and corresponding relationship strength is set. At least one relationship path between a first user node and a second user node is determined according to the user relationship graph; and a corresponding candidate relation data is determined based on a comparison result between a plurality of relation strengths corresponding to a plurality of relation edges included in each relation path.
JP-2022082523 to Song (hereinafter “Song”). Song discloses providing information about machine learning based similar items by receiving information about a target item; generating a target vector based on a character string corresponding to the information about the target item using a machine learning model; checking at least one vector set respectively corresponding to a plurality of items derived through the machine learning model; and providing information about at least one item corresponding to at least one vector having a similarity value with the generated target vector greater than or equal to a preset critical value in the at least one vector set.
“Incorporating Similarity Measures to Optimize Graph Convolutional Neural Networks for Product Recommendation” by Shafqat et al., dated 2021 (hereinafter “Shafqat”). Shafqat discloses a model that incorporates measures of similarity between two different nodes, and these similarity measures help to sample the neighbors beforehand. The similarity is estimated based on their interaction probability distribution with other nodes. A KL divergence is used on different probability distributions to find the distance between them.
None of the prior references cited herein teach or suggest the above discussed limitations in combination with the other claim language in the independent claims.
Prior Art
The following prior art, made of record and not relied upon, is considered pertinent to Applicant’s disclosure:
“PersonaSAGE: A Multi-Persona Graph Neural Network” by Choudhary et al., dated 2023 (hereinafter “Chouduary”). Chouduary discloses develop a persona based graph neural network framework called PersonaSAGE that learns multiple persona-based embeddings for each node in the graph.
Conclusion
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 Rupangini Singh whose telephone number is 571-270-0192. The examiner can normally be reached on Monday – Friday, 9:30 AM – 6:30 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, Shannon Campbell can be reached on Monday – Friday at (571) 272-5587. 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.
/RUPANGINI SINGH/
Primary Examiner, Art Unit 3628
Prosecution Insights