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 .
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1, 3, 5-12, 16-17 and 19-20, is/are rejected under 35 U.S.C. 103 as being unpatentable over Deffayet et al., (U.S. Pub. No. U.S. Pub. No. 2024/0289631 A1) in view of Wang et al., (U.S. Pub. No. 2025/0247582 A1).
As per claim 1, Deffayet teaches a computer-implemented method, comprising: training an end-to-end generative slate recommendation model, wherein training the end-to-end generative slate recommendation model includes (abstract, [0009], [0042], fig. 3-5); fine-tuning tuning the sequence model using the training dataset and a plurality of slate information, to generate a fine-tuned sequence model configured to determine recommended content items further based at least in part on relationships between the sequence of recommended content items for populating a slate recommendation for a user ([0055], [0073], [0141], and fig. 2); training, using the fine-tuned sequence model and a plurality of feedback information, at least one reward model based at least in part on an objective (abstract, [0006], [0009],[0031], [0036], [0141], “…the recommender system 108 includes a reinforcement learning (RL) model that is trained to take an action (i.e., recommend a slate 110) for each turn) to optimize a reward”; fig. 1); and fine-tuning the fine-tuned sequence model using the at least one reward model and at least one of a reinforcement learning technique or a direct preference optimization technique to generate the end-to-end generative slate recommendation model (abstract, [0006], [0009], [0026], [0031], [0033-0036], figs. 1-5).
Deffayet does not explicitly disclose accessing a training dataset including a plurality of contextual training data; training, using the training dataset, a sequence model configured to determine a sequence of recommended content based on an input user history and an input contextual information.
However, Wang teaches accessing a training dataset including a plurality of contextual training data (fig. 3A-3C, fig. 4A; 5A abstract, [0045]); using the training dataset, a sequence model configured to determine a sequence of recommended content based on an input user history and an input contextual information (abstract, fig. 2A-5A; [0029], [0030], [0032-0035]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 3, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 1. In addition, Deffayet teaches wherein the fine-tuned sequence model biases the end-to-end generative slate recommendation model to determine slate recommendations base at least in part on the objective ([0077]).
As per claim 5, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 1. Deffayet does not explicitly disclose wherein fine-tuning of the sequence model includes training the sequence mode to imitate a recommendation system employing one or more machine learning models.
However, Wang teaches wherein fine-tuning of the sequence model includes training the sequence model to imitate a recommendation system employing one or more machine learning models (fig. 5A, [0044-0045]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 6, Deffayet teaches a computer-implemented method, comprising: receiving a request for a slate of content for a user (fig. 1 el. 102, fig. 2; and [0030-0033]); processing, using a trained generative slate recommendation model ([0024], [0026], [0029-0037] and fig. 1 el. 108), a sequence of user interactions associated with the user and a contextual information to determine a slate recommendation for the user ([0024], [0031], [0045]; figs. 1-2), wherein: the slate recommendation includes a recommend sequence of content items (fig. 1, output from 108; fig. 2); and the trained generative recommendation mode was trained based at least in part on an initial model and a reward model ([0031]); and causing the slate recommendation to be presented on a client device associated with the user ([0031]; [0137] and fig. 10, the recommended slate 110 of items may be transmitted to a user device (e.g., a web browser, application (app), etc.) provided for display on display (e.g., screen), printed, provided for presenting by a virtual assistant or bot, etc.).
As per claim 7, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 6. In addition, Deffayet teaches wherein the sequence of user interactions includes a sequence of content items with which the user interacted ([0034]; “… the dataset may be generated offline, generated based on prior interactions with the user, …”).
As per claim 8, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 7. In addition, Deffayet teaches wherein at least one of the sequence of content items or the recommended sequence of items includes content items of more than one content item type ([0030], “…the retriever 104 retrieves a collection of items 106 (e.g., unique identifiers (IDs) respectively associated with the items, or any combination of the items and the IDs). The items in the collection may include, for instance, media items, documents, tokens, news items, terms, e-commerce items, or a combination thereof”).
As per claim 9, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 7. In addition, Deffayet teaches everything as claimed above, see claim 7. In addition, Deffayet teaches wherein the sequence of content items are represented as a sequence of embeddings encoded features of content items included in the sequence of content items ([0033], [0065], [0141], and fig. 5).
As per claim 10, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 7. Deffayet does not explicitly disclose wherein the sequence of content items includes a first content item encoded as a respective embedding that is configured to dynamically change based on user interactions with the first dynamic content item.
However, Wang teaches wherein the sequence of content items includes a first content item encoded as a respective embedding that is configured to dynamically change based on user interactions with the first dynamic content item ([0031], [0033], [0040-0041], [0050], fig.2A-4C).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 11, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 7. Deffayet does not explicitly disclose wherein the plurality of contextual information includes at least one of: a type of request for the slate of content; a time of the request for the slate of content; a device type; a device display type; or a device display orientation.
However, Wang teaches wherein the plurality of contextual information includes at least one of: a type of request for the slate of content; a time of the request for the slate of content; a device type; a device display type; or a device display orientation ([0030], [0035], fig. 3A).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 12, Deffayet (modified by Wang)as a whole teaches everything as claimed above, see claim 6. In addition, Deffayet teaches wherein the type of request for the slate of content includes at least one of: a query; a request to access a homepage; a shopping session; or a request for recommended content ([0137]).
As per claim 16, Deffayet teaches a computing system, comprising: one or more processors (abstract; [0010], [0133] fig. 10);and a memory storing program instructions that, when executed by the one or more processors, cause the one or more processor (abstract, and [0133] and fig. 10) to at least: obtain a slate recommendation model configured to determine slates of recommended content items (abstract, [0132], figs. 1-5); determine at least one objective for a reward model ([0031], [0141]); generate a reward model to determine a reward for slates of recommended content items determined by the slate recommendation model based at least in part on the at least one objective ([0009], [0024], [0026], [0031], and fig. 1-5); optimize, based at least in part on the reward model, the slate recommendation model to generate an optimized slate recommendation model configured to determine slate recommendations based at least in part on the at least one objective ([0031] and fig.1-5); receive a request for a slate content for a user ([0029] and fig. 1); and return the user slate recommendation (fig. 1-5 and [0032-0035]). Deffayet does not explicitly disclose to process, using the optimized slate recommendation model, a sequence of a user interactions associated with the user and a plurality of contextual information associated with the request to determine a user slate recommendation.
However, Wang teaches process, using the optimized slate recommendation model, a sequence of a user interactions associated with the user and a plurality of contextual information associated with the request to determine a user slate recommendation (abstract, fig. 2A-5A; [0029], [0030], [0032-0035]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 17, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 16. In addition, Deffayet teaches everything as claimed above, see claim 16. In addition, Deffayet teaches wherein optimizing the slate recommendation model includes employing at least one of a reinforcement learning technique or a direct preference optimization technique ((abstract, [0006], [0009], [0026], [0031], [0033-0036], figs. 1-5).
As per claim 19, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 16. Deffayet does not explicitly disclose wherein the plurality of contextual information includes a type of request for the slate of content; and the type of request for the slate content includes at least one of: a query; a request to access a homepage; a shopping session; a request to push content; or a request for recommended content.
However, Wang teaches wherein the plurality of contextual information includes a type of request for the slate content (fig. 3A); and the type of request for the slate content included at least one of a query; a request to access a homepage; a shopping session; a request to push content; or a request for recommend content (fig. 2A, 3A-3C; [0030], [0035]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Wang with Deffayet for the benefit of providing improved content recommendations.
As per claim 20, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 16. In addition, Deffayet teaches wherein the user slate recommendations includes a plurality of representations encoding a sequence of content items forming the user slate recommendations (figs. 1-4).
Claim(s) 2, 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Deffayet et al., (U.S. Pub. No. 2024/0289631 A1) in view of Wang et al., (U.S. Pub. No. 2025/0247582 A1) and further in view of Sharpe et al., (U.S. Pub. No. 2024/0394553 A1).
As per claim 2, Deffayet teaches everything as claimed above, see claim 1. In addition, Deffayet teaches wherein the at least one reward model is model is configured to determine a reward associated with the slate recommendation ([0031], [0141-0142]). Deffayet does not explicitly disclose the reward includes a scalar value representing a quality of the slate recommendation based at least in part on the objective.
However, Sharpe teaches the reward includes a scalar value representing a quality of the slate recommendation based at least in part on the objective ([0037]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Sharpe with Deffayet (modified by Wang) for the benefit of providing effective training of agents to make recommendations and decisions.
As per claim 15, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 16. Deffayet does not explicitly disclose wherein training a reward model is based at least in part on a plurality of objectives.
However, Sharpe teaches wherein training reward model is based at least in part on a plurality of objectives ([0018]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Sharpe with Deffayet (modified by Wang) for the benefit of providing effective training of agents to make recommendations and decisions.
Claim(s) 13-14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Deffayet et al., (U.S. Pub. No. 2024/0289631 A1) in view of Wang et al., (U.S. Pub. No. 2025/0247582 A1) and further in view of Cartoon et al., (U.S. Pub. No. 2016/0299906 A1).
As per claim 13, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 6. Deffayet does not explicitly disclose wherein determination of content items in the recommended sequence of content items is based at least in part on preceding content items in the recommended sequence of content items.
However, Cartoon teaches wherein determination of content items in the recommended sequence of content items is based at least in part on preceding content items in the recommend sequence of content items ([0054]; “… if the recommended content attribute sequence identifies a happy content item, followed by a bouncy content item, followed by another happy content item, content item recommendation module 135 can generate the recommended content item sequence by selecting two happy content items and one bouncy content item and then ordering the selected content items according to the recommended content attribute sequence (e.g., happy-bouncy-happy).”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Cartoon with Deffayet (modified by Wang) for the benefit of providing improved recommendations for sequence of content items.
As per claim 14, Deffayet (modified by Wang) as a whole teaches everything as claimed above, see claim 6. Deffayet does not explicitly disclose wherein determination of content items in the recommend sequence of content items is not based on subsequent content items in the recommended sequence of content items.
However, Cartoon teaches wherein determination of content items in the recommend sequence of content items is not based on subsequent content items in the recommended sequence of content items ([0054]; “… if the recommended content attribute sequence identifies a happy content item, followed by a bouncy content item, followed by another happy content item, content item recommendation module 135 can generate the recommended content item sequence by selecting two happy content items and one bouncy content item and then ordering the selected content items according to the recommended content attribute sequence (e.g., happy-bouncy-happy).”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Cartoon with Deffayet (modified by Wang) for the benefit of providing improved recommendations for sequence of content items.
Allowable Subject Matter
Claims 4 and 18 are 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.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Xiao et al., (U.S. Pub. No. 2024/0420013 A1), “Debiased Training of Machine Learning Systems”
Kadioglu et al., (U.S. Pub. No. 2024/0232652 A1), “Digital Content Classification and Recommendation Using Constraint-Based Predictive Machine Learning”
Contact
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JESSICA PRINCE whose telephone number is (571)270-1821. The examiner can normally be reached M-F 7:30-3:30 P.M..
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, Jamie Atala can be reached at 571-272-7384. 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.
JESSICA PRINCE
Examiner
Art Unit 2486
/JESSICA M PRINCE/Primary Examiner, Art Unit 2486
Prosecution Insights