DETAILED ACTION
This Final Office Action is in response Applicant communication filed on
12/22/2025. In Applicant’s amendment, claims 1-5, 9, 11-14, and 20 were amended. Claims 7, 8, 17, and 18 are cancelled. Claims 1-6, 9-16, 19, and 20 are currently pending and have been rejected as follows.
Response to Amendments
Rejections under 35 USC 101 are maintained. Applicant’s amendments necessitated new grounds of rejection under 35 USC 103.
Response to Arguments
Applicant’s 35 USC 101 rebuttal arguments and amendments have been fully considered but they are not persuasive to overcome the rejection.
Applicant argues on p. 12-13 that a base “neural network” for recommending a source of a gift converted into one that is specific to the gift giver and recipient is a technical improvement in the training of a neural network because a neural network trained for an interaction between two different people is a clear technical improvement not achievable with conventional training. Examiner respectfully disagrees.
The claim is recited at a result-based functional level. The alleged “converting” of the base neural network into a user pair specific one is done so at a high level without concrete technical details beyond generic retraining on feedback and browsing data. The alleged integration is better recommendation relevance, not an improvement to how machine learning models are trained. The user devices do no more than generic interface input/output in a recommendation workflow, amounting to insignificant extra-solution activity. Applying an abstract idea in a neural networking environment at a high level amounts to mere instructions to apply the abstract idea to a computer or technical environment.
Examiner has updated the rejection removing the mathematical concepts grouping from the abstract idea. The primary abstract idea of certain methods of organizing human activity remain with some steps including mental processes. The claims remain rejected without integrate the judicial exception into a practical application and without reciting significantly more.
Applicant's prior art arguments have been fully considered but they are moot in light of the newly cited Centitas and Linden references.
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 1-6, 9-16, 19, and 20 are clearly drawn to at least one of the four categories of patent eligible subject matter recited in 35 U.S.C. 101 (method, system, and non-transitory computer readable medium). Claims 1-6, 9-16, 19, and 20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without integrating the abstract idea into a practical application or amounting to significantly more than the abstract idea.
Regarding Step 1 of the 2019 Revised Patent Subject Matter Eligibility Guidance (‘2019 PEG”), Claims 1-6, 9-10 are directed toward the statutory category of a process (reciting a “method”). Claims 11-16, 19 are directed toward the statutory category of a machine (reciting a “system”). Claim 20 is directed toward the statutory category of an article of manufacturer (reciting a “non-transitory computer readable medium”).
Regarding Step 2A, prong 1 of the 2019 PEG, Claims 1, 11 and 20 are directed to an abstract idea by reciting obtaining, …, historical transaction data associated with a first user; determining, … and based on the historical transaction data, at least one party with whom the first user had a threshold quantity of transactions; identifying, at least one potential source of a gift for the first user by comparing the at least one party to a plurality of parties identified …; determining, …, a respective target metric for each potential source by inputting the historical transaction data into a neural network trained on training transaction data of one or more other users and training sources of gifts given to the one or more other users; causing … associated with the first user to output an identification of the at least one potential source, with reference to the respective target metric of each potential source; receiving, …, feedback regarding the identification; re-training the neural network based on the feedback to convert the neural network into a user-specific neural network for the first user; using the re-trained neural network, refining the identification of the at least one potential source; causing, … associated with a second user to output the refined identification of the at least one potential source; receiving an indication … that the second user has visited a website associated with one or more sources; and further re-training the user-specific neural network based on the indication to convert the user-specific neural network into a neural network specific to the first user and the second user (Example Claim 1).
The claims are considered abstract because these steps recite certain methods of organizing human activity like advertising, marketing or sales activities or behaviors and mental processes. The claims focus on recommending gift sources for one person to another person based on transaction history and then personalizing that recommendation using feedback and browsing behavior. Applicant’s disclosure does not recite a particular problem the claimed steps aim to solve, however, it is understood that the claimed steps aim to access first user interaction records related to third parties to train a model to identify and provide gifting options to a second user associated with the first user (Applicant’s Specification, [0018]-[0023]). By this evidence, the claims recite a type of certain methods of organizing human activity like advertising, marketing or sales activities or behaviors and mental processes common to judicial exception to patent-eligibility. By preponderance, the claims recite an abstract idea (e.g., a “system and method” for training a machine learning model to identify targets).
Regarding Step 2A, prong 2 of the 2019 PEG, the judicial exception is not integrated into a practical application because the claims (the judicial exception and the additional elements such as by at least one processor; a target database; a user device; a second user device) are not an improvement to a computer or a technology, the claims do not apply the judicial exception with a particular machine, the claims do not effect a transformation or reduction of a particular article to a different state or thing nor do the claims apply the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment such that the claims as a whole is more than a drafting effort designed to monopolize the exception (see MPEP §§ 2106.05(a-c, e)).
Dependent claims 2-6, 12-16, and 19 do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the limitations recite mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea ‐ see MPEP 2106.05(f).
Regarding Step 2B of the 2019 PEG, the additional elements have been considered above in Step 2A Prong 2. The claim limitations do not amount to significantly more than the judicial exception because they are directed to limitations referenced in MPEP 2106.05I.A. that are not enough to qualify as significantly more when recited in a claim with an abstract idea because the limitations recite mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea ‐ see MPEP
2106.05(f).
Applicant's claims mimic conventional, routine, and generic computing by their similarity to other concepts already deemed routine, generic, and conventional [Berkheimer Memorandum, Page 4, item 2] by the following [MPEP § 2106.05(d) Part (II)]. The claims recite steps like: “Receiving or transmitting data over a network, e.g., using the Internet to gather data,” Symantec, “Performing repetitive calculations,” Flook, and “storing and retrieving information in memory,” Versata Dev. Group, Inc. v. SAP Am., Inc. (citations omitted), by performing steps of “obtaining” data, “determining” at least one party, “identifying” at least one potential source of a gift, “determining” a target metric, “causing” output of a potential source, “ receiving” feedback, and “re-training” the machine learning model, “using” the re-trained model, “causing” output of the refined identification of the potential source, “receiving” an indication of browsing history, and “re-training” the user-specific neural network (Example Claim 1).
By the above, the claimed computing “call[s] for performance of the claimed information collection, analysis, and display functions ‘on a set of generic computer components' and display devices” [Elec. Power Group, 830 F.3d at 1355] operating in a “normal, expected manner” [DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d at 1245, 1258 (Fed. Cir. 2014)].
Conclusively, Applicant's invention is patent-ineligible. When viewed both individually and as a whole, Claims 1-6, 9-16, 19, and 20 are directed toward an abstract idea without integration into a practical application and lacking an inventive concept.
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.
Claims 1-6, 9-16, 19, and 20 are rejected under 35 USC 103 as being unpatentable over the teachings of
Chang et al., US 20160253737 A1, hereinafter Chang, in view of
Pavlidis et al., US 20140089327 A1, hereinafter Pavlidis, in view of
Centitas et al., US 20210398193 A1, hereinafter Centitas, in view of
Linden et al., US 20130332308 A1, hereinafter Linden. As per,
Claims 1, 11, 20
Chang teaches
A computer-implemented method for training a machine learning model, the method comprising: /
A system for training a machine learning model, the system comprising: a memory storing instructions; and a processor operatively connected to the memory and configured to execute the instructions to perform operations including: /
A non-transitory computer readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations for training a machine learning model, the operations comprising: (Chang [0027]; [0114])
obtaining, by at least one processor, historical transaction data associated with a first user; (Chang [0030] “The SEED Merchants are used to identify consumers that have past transactions with the SEED merchants in order to generate a list of identified consumers;” [0102] ““record of charge” or “ROC” may comprise a record of a transaction or charge by a consumer with a particular merchant;” [0108] “a consumer profile may include online tracking cookie data, web beacon data, web tracking data, web packet trace data, digital fingerprint data, clickstream data, purchase or transaction history data” note the use of consumers past transactions and transaction database records)
determining, by the at least one processor and based on the historical transaction data, at least one party with whom the first user had a threshold quantity of transactions; (Chang [0030]; [0043] “The list of active merchants may be further filtered by deriving whether the merchant conducts sufficient consumer activity with distinct consumers. For example and with reference to FIG. 3B, active merchants with a limited record of distinct spending consumers, such as less than 11, are removed from the merchant list” note the threshold quantity of transactions based filtering to identify at least one party)
identifying, at least one potential source of a gift for the first user by comparing the at least one party to a plurality of parties identified in a target database; (Chang [0033] “analysis cycle 100 may include scoring the merchants in the scored merchant list based on a number of distinct merchants and over-indexing” note the scoring of merchants in relation to each other)
[…] inputting the historical transaction data […]; (Chang [0030]; [0043])
[…];
[…];
[…];
[…];
[…];
[…];
[…].
Chang does not explicitly teach, Pavlidis however in the analogous art of generating recommendations teaches
causing a user device associated with the first user to output an identification of the at least one potential source, with reference to the respective target metric of each potential source; (Pavlidis fig. 10 “Present Gift Recommendations)
receiving, via the user device, feedback regarding the identification; (Pavlidis fig. 6 noting the presenting of a product in interest contexts, monitoring the interaction, and updating the interest score of the product; [0067] “The method 600 of FIG. 6 may be used to define and refine the interests and attributes associated with a product” note the refining of interests and attributes based on the monitored interaction)
Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify Chang’s recommendations to include user device output recommendations and feedback in view of Pavlidis in an effort to resolve the ambiguity in gift giving (see Pavlidis ¶ [0021] & MPEP 2143G).
Chang / Pavlidis do not explicitly teach, Centitas however in the analogous art of generating recommendations teaches
determining, by the at least one processor, a respective target metric for each potential source by […] into a neural network trained on training transaction data of one or more other users and training sources of […]; (Centitas [0007] “a trained neural contextual bandit model) to model user-item interactions, adaptively learn the underlying representations of users or items (e.g., user's preference or item's characteristic) using exploration and exploitation strategies, and provide an expected reward (or payoff) representing a user's preference in connection with each of a number of items” note the expected reward/payoff corresponding to a respective target metric)
re-training the neural network based on the feedback to convert the neural network into a user-specific neural network for the first user; (Centitas [0008] “n each round (or trial), the GRC updates user and item feature representations based on user-item interactions;” [0014] “The initial representation determined for the new user can be updated and supplemented using additional information;” [0113] “in an exploration mode, the neural contextual bandit framework 606 uses the user’s feedback (e.g., positive, negative, unobserved, etc.) for training its neural network model using the knowledge collected about the user” noting the iterative retraining/updating based on collected feedback for a particular user)
using the re-trained neural network, refining the identification of the at least one potential source; (Centitas [0008] “In each round (or trial), the GRC updates user and item feature representations based on user-item interactions, and uses the updated user and item feature representations to determine an expected reward for each item in the current round;” [0105] “the user and item representations (together with their interrelationships in feature space) determined at steps 406 and 408 are used to determine an estimated reward (e.g., an estimated user preference and likelihood of positive user feedback) for each item of the plurality of items;” [0106] “the estimated reward determined for each candidate item is then used to select one or more candidate items for inclusion in the item recommendation” note once feedback is incorporated, the updated model is used to generate improved expected rewards and select refined recommendations)
further re-training the user-specific neural network based on the indication to convert the user-specific neural network into […]. (Centitas [0008] “In each round (or trial), the GRC updates user and item feature representations based on user-item interactions, and uses the updated user and item feature representations to determine an expected reward for each item in the current round” noting the retraining based on feedback)
Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify Chang’s recommendations and Pavlidis’s feedback to include target metrics, user specific models, and re-training the user specific model in view of Centitas in an effort to improve computer generated results (see Centitas ¶ [0028] & MPEP 2143G).
Chang / Pavlidis / Centitas do not explicitly teach, Linden however in the analogous art of generating recommendations teaches
[…] gifts given to the one or more other users; (Linden [0079] “rank the selected gifts, such as according to estimated interest of the recipient, estimated relevance to the sender, price, complementary correlation with a gift sent to the recipient by another user, or perceived recipient need” note the gifts)
causing, by the at least one processor, a second user device associated with a second user to output the refined identification of the at least one potential source; (Linden [0043] “The suggestion engine 824 can subsequently recommend the selected gift to the sender and facilitate sender purchase of the product for the recipient.” Note the sender corresponding to a second user device and receiving the recommended selected gift for purchase for the recipient)
receiving an indication from the second user device that the second user has visited a website associated with one or more sources; and (Linden [0040] “The action log 832 records data about these users, including webpage viewing histories, advertisements that were engaged, purchases made” noting the external browsing action logged for a user)
[…] a neural network specific to the first user and the second user. (Linden [0043] “The suggestion engine 824 can further calculate an affinity between a user who is a (potential) recipient and a second user who is a (potential) sender, such as based on interactions between the users including messages, posts, and/or other communications between the users within the social networking system, and select the second user as the sender based on the affinity between the users. The suggestion engine 824 can subsequently recommend the selected gift to the sender and facilitate sender purchase of the product for the recipient.” Note the relationship modeled between the sender and recipient)
Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify Chang’s recommendations, Pavlidis’s feedback, and Centitas’s refined model to include a second user receiving notifications and a modeled relationship between the gift giver and gift receiver in view of Linden in an effort to recommend relevant gifts meeting determined criteria (see Linden ¶ [0072] & MPEP 2143G).
Claims 2, 12
Chang teaches
wherein determining the respective target metric further comprises: obtaining information related to each source from one or more websites associated with each source in the target database; and (Chang [0103] “A “channel” may include any system or method for delivering content … a channel may comprise a website;” [0068] “recommendations or offers may be ranked using any suitable factors. For example, a persona may comprise a plurality of recommendations. These recommendations may be ranked and/or presented to a consumer using any suitable factors, based on the channel being accessed by the consumer … the system presenting the personas may also monitor activity of a user associated with the channel” note the website data)
inputting the information related to each target to the machine learning model along with the historical transaction data to determine the respective target metric. (Chang [0030]; [0043]; [0068] “the system presenting the personas may also monitor activity of a user associated with the channel, such that the system understands that the particular consumer's tendencies in response to the consumer accessing the channel. For example, if a threshold percentage (e.g., 90%) of consumer activity in a particular channel is related to a particular item (e.g., restaurant reconditions, “buy one/get one” offers, item discounts, and/or the like), the system may adjust the ranking of identified offers for a particular persona based on the channel” note the system inputting the website and interaction data for the user to adjust the recommendations/offers)
Claims 3, 13
Chang / Centitas / Linden do not explicitly teach, Pavlidis however in the analogous art of generating recommendations teaches
wherein the information related to each target comprises information related to providing gifts. (Pavlidis [0092] “ products may also be ranked 1016 or filtered according to giftability. Giftability may indicate the appropriateness for a product to be given as a gift”)
The motivation/rationale to combine Chang / Centitas / Linden with Pavlidis persists.
Claims 4, 14
Chang teaches
wherein determining the at least one party comprises: determining whether a classification code associated with each of the transactions in the historical transaction data is of a predetermined type, the classification code indicating a classification of a party with whom each of the transactions occurred. (Chang [0030] “The initial seed, or search input, may include an industry code”)
Claims 5, 15
Chang / Centitas / Linden do not explicitly teach, Pavlidis however in the analogous art of generating recommendations teaches
wherein the feedback regarding the identification comprises an indication that the first user rejects the at least one potential source. (Pavlidis [0046] “Where positive or negative sentiment has already been noted in the profile, the sentiment may be augmented or decremented, i.e., by adjusting a numerical value, according to whether the identified sentiment for the entity agrees of disagrees, respectively, with the previously recorded sentiment for the entity;” fig. 6 noting the presenting of a product in interest contexts, monitoring the interaction, and updating the interest score of the product; [0067] “The method 600 of FIG. 6 may be used to define and refine the interests and attributes associated with a product”)
The motivation/rationale to combine Chang / Centitas / Linden with Pavlidis persists.
Claims 6, 16
Chang teaches
wherein the respective target metric indicates an ease with which a second user may acquire an item intended for the first user from the at least one party. (Chang [0009] “the system may rank the first redemption channel higher than the second redemption channel to promote benefits that are easy for the user to receive” note the ease of acquisition considered in the ranking)
Claims 9, 19
Chang teaches
further comprising: determining that the one or more sources in the indication do not match any of the plurality of parties identified in the target database; and adding the at least one party to the target database. (Chang fig. 2; [0034] “At Step 2B, other merchants visited by the identified consumers are retrieved from a transaction database;” [0038] “Step 2D illustrates a step of processing an updated merchant list” note the addition of a merchant that was not previously in the list of merchants)
Claim 10
Chang teaches
further comprising: authenticating, by at least one processor, the user device associated with the first user. (Chang [0109] “a consumer may specify that the consumer's profile may be based upon data associated with the consumer's transaction history;” [0110] “Phrases and terms similar to “account,” “transaction account,” “account,” “account number,” “account code,” and/or “consumer account” may include any account that may be used to facilitate a financial transaction. These accounts may include any device, code (e.g., one or more of an authorization/access code, personal identification number (“PIN”), Internet code, other identification code, and/or the like), number, letter, symbol, digital certificate, smart chip, digital signal, analog signal, biometric or other identifier/indicia suitably configured to allow the consumer to access, interact with or communicate with the system”)
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. US 20180260693 A1; WO 2017/123415 A1; Forte et al., Learning human behavioral profiles in a cyber environment, 2010.
THIS ACTION IS MADE FINAL. 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 extension fee 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 MOHAMED EL-BATHY whose telephone number is (571)270-5847. The examiner can normally be reached on M-F 8AM-4:30PM.
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, PATRICIA MUNSON can be reached on (571) 270-5396. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MOHAMED N EL-BATHY/Primary Examiner, Art Unit 3624
