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 § 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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more. The claims recite an abstract idea. This judicial exception is not integrated into a practical application. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Under Step 1 of the eligibility analysis the claims are directed to statutory categories. MPEP 2106.03. Specifically, the method, as claimed in claims 1-7, is directed to the process. Additionally, the medium, as claimed in claims 8-14, is directed to an apparatus. Finally, the machine-storage medium, as claimed in claims 15-20, is directed to a machine.
While the claims fall within statutory categories, under Step 2A, Prong 1 of the eligibility analysis (MPEP 2106.04), the claimed invention recites the abstract idea of providing next item recommendation to a user. Specifically, representative claim 1 recites the abstract idea of:
Receiving current event information characterizing a current event related to a communication session involving a user and the user’s interactions;
Retrieving historic sequence data related to a plurality of past user interactions, wherein the historic sequence data includes performance data with respect to content recommended to the users and displayed;
Creating a task sentence, wherein the task sentence has multiple tokens, each of which corresponds to a task;
Generating data, respectively, for the current event information, the historic sequence data, and the task sentence;
predicting, based on the data, a next item to be recommended to the user; and
presenting the next item to the user.
Under Step 2A, Prong 1 of the eligibility analysis, it is necessary to evaluate whether the claim recites a judicial exception by referring to subject matter groupings enumerated in MPEP 2106.04(a). The abstract idea identified above is considered to be a certain method of organizing human activity. Certain methods of organizing human activity include “fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions).”” MPEP 2106.04(a)(2)(II). In this case, the abstract idea recited in representative claim 1 is a certain method of organizing human activity because receiving current event information, retrieving historic sequence data, creating a task sentence, predicting a next time to recommend, and presenting the next item is a commercial or legal interaction because it is a sales activity and/or relates to business relations. Thus, representative claim 1 recites an abstract idea.
Under Step 2A, Prong 2 of the eligibility analysis, if it is determined that the claims recite a judicial exception, it is then necessary to evaluate whether the claims recite additional elements that integrate the judicial exception into a practical application of that exception. MPEP 2106.04(d). The courts have identified limitations that did not integrate a judicial exception into a practical application include limitations merely 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, as discussed in MPEP 2106.05(f). MPEP 2106.04(d). In this case, representative claim 1 includes additional elements such as a plurality of user interfaces, a mixture of expert prediction model trained via machine learning with multi-task learning. Although reciting such additional elements, the additional elements do not integrate the abstract idea into a practical application because they merely amount to no more than an instruction to apply the abstract idea using a generic computer or merely use a computer as a tool to perform the abstract idea. These additional elements are described at a high level in Applicant's specification without any meaningful detail about their structure or configuration. Similar to the limitations of Alice, representative claim 1 merely recites a commonplace business method (i.e., providing item recommendation) being applied on a general-purpose computer. See MPEP 2106.05(f). Thus, the claimed additional elements are merely generic elements and the implementation of the elements merely amounts to no more than an instruction to apply the abstract idea using a generic computer/interface. Since the additional elements merely include instructions to implement the abstract idea on a generic computer/interface or merely use a generic computer model as a tool to perform an abstract idea, the abstract idea has not been integrated into a practical application.
Under Step 2B of the eligibility analysis, if it is determined that the claims recite a judicial exception that is not integrated into a practical application of that exception, it is then necessary to evaluate the additional elements individually and in combination to determine whether they provide an inventive concept (i.e., whether the additional elements amount to significantly more than the exception itself). MPEP 2106.05. In this case, as noted above, the additional elements recited in independent claim 1 are recited and described in a generic manner merely amount to no more than an instruction to apply the abstract idea using a generic computer/interface or merely use a generic computer model as a tool to perform an abstract idea.
Even when considered as an ordered combination, the additional elements of representative claim 1 do not add anything that is not already present when they considered individually. In Alice, the court considered the additional elements “as an ordered combination,” and determined that “the computer components...‘ad[d] nothing. ..that is not already present when the steps are considered separately’... [and] [v]iewed as a whole...[the] claims simply recite intermediated settlement as performed by a generic computer.” Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 217, (2014) (citing Mayo, 566 U.S. at 79, 101 USPQ2d at 1972). Similarly, when viewed as a whole, representative claim 1 simply conveys the abstract idea itself facilitated by generic computing components. Therefore, under Step 2B of the Alice/Mayo test, there are no meaningful limitations in representative claim 1 that transforms the judicial exception into a patent eligible application such that the claims amount to significantly more than the judicial exception itself.
As such, representative claim 1 is ineligible.
Dependent Claims 2-7 do not aid in the eligibility of independent claim 1. For example, claims 2-9 merely further define the abstract limitations of claim 1. Dependent claims 2-7 do not recite additional elements supplemental those recited in claim 1. Therefore, the additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea for the reasons described above with respect to claim 1. Thus, dependent claims 2-7 are also ineligible.
Independent claims 8 and 15 recite the same abstract idea represented in representative claim 1. In addition to the additional elements of claim 1, Independent Claim 8 recites the additional elements of a machine readable and non-transitory medium and machine. In addition to the additional elements of claim 1, Independent Claim 19 recites the additional elements of a system comprising a contextual information collector implemented by a processor, a next item recommendation engine implemented by a processor. The additional elements in Independent claims 8 and 15 do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea for the reasons described above with respect to claim 1.
Similarly, the dependent claims 9-14 and 16-20 do not recite additional elements supplemental those recited in claims 2-7. Therefore, the additional elements to not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea for the reasons described above with respect to claims 2-7, respectively.
Thus, dependent claims 9-14 and 16-20 are also ineligible.
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.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Javari et al. (US 2024/0220810) in view of Ullrich et al. (US 2024/0419246), and further in view of Li et al. (US 2022/0019878).
Regarding Claims 1, 8, and 15, Javari discloses a method for recommendation, comprising: (See at least paragraph [0003], [0013])
Receiving current event information characterizing a current event related to a communication session involving a user and the user’s interactions with a user interface (UI) (See at least paragraph [0049] disclosing current event involving user’s interactions with interface as a sequence that user selects, clicks, navigates, scrolls, adds to cart, purchases, etc., [0086] disclosing receiving user actions through a user interface, [0087] disclosing input as sequence of user actions);
Retrieving historic sequence data related to a plurality of UIs and past user interactions directed thereto, wherein the historic sequence data includes performance data with respect to content recommended to the users and displayed on the plurality of UIs (See at least paragraph [0003] disclosing learn user preferences form historical user interactions and then recommend items aligned with learned preferences of the user, [011] disclosing users historical actions represented as a sequence over time, [0013], [0026] disclosing historical interactions such as items that frequently occur together, etc., [0027]);
Generating embeddings, respectively, for the current event information, the historic sequence data, and the task sentence (See at least Abstract, paragraph [011] embedding with use sequence, user historical interactions, etc., [0014]-[0015], [0032]-[033] disclosing embeddings encoded with sequential data, context, self-attention, items/metadata, etc., [0050] disclosing sequence information of new user action, prior interactions, etc. and inputting sequence of user actions to metadata embedding module, etc., [0079]);
Predicting, based on the embeddings vi a prediction model, a next item to be recommended to the user on the UI, wherein the prediction model is previously trained via machine learning with multi-task learning (See at least paragraph [0011] disclosing next item prediction to recommend items based on embeddings, [0015] disclosing integrating item meta feature and leveraging multi-task learning, [0032] disclosing recommendation system obtains given user sequence that goes through embeddings and inputs into trained model to make predictions, Fig. 2b, 3, 4);
Presenting the next item to the user on the UI (See at least paragraph [0003], Fig. 1, [0022] using machine learning model and website to recommend next document to user, [0033], Fig. 3 & [0045] disclosing providing recommendation to user, [0053] disclosing displaying through electronic interface the recommendation for an item based on predicted next user actions).
Javari does not expressly provide for creating a task sentence with respect to the UI, wherein the task sentence has multiple tokens, each of which correspond to a task. However, Ulrich discloses creating a task sentence with respect to the UI, wherein the task sentence has multiple tokens, each of which correspond to a task (See at least paragraph [0066] disclosing using event data to be communicated in tokenized forms, [0067], [0068] disclosing user input prompt tokens, set of tokens, etc., [0070], [0085], [0086] disclosing token sequence generated as multimodal output, [0090], [0091] disclosing user input prompt tokens which represent users desired configuration, etc., [0101], [0105], [0116] disclosing feedback tokens, [0122] disclosing user response and actions may be recorded and combined with tokens, [0123], [0128], [0359] disclosing string of tokens form prompt composer and generating feedback related to the specific task, [0429]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the tokens, as taught by Ulrich, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including better prompts for the model based on information other than just natural language provided by a user. See at least paragraph [0005], [0085].
Neither Javari nor Ulrich expressly provides for a mixture of expert (MoE) prediction model. However, Li discloses a mixture of expert (MoE) prediction model (See at least paragraph [0006], [0007], [0026], [0030], [0069]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the MoE, as taught by Li, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including ability for large volumes to be handled and optimized ranking objectives where parameters are automatically learned to share across potentially conflicting objectives and bring improvements to full recommendation system. See at least paragraph [0026], [0030], [0069].
Regarding Claims 2, 9, and 16, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 1, 8, and 15. Javari does not expressly provide for wherein the MoE prediction model comprises a plurality of experts that are trained via the multi-task learning to gain respective expertise. However, Li discloses wherein the MoE prediction model comprises a plurality of experts that are trained via the multi-task learning to gain respective expertise (See at least Claim 4, paragraph [0026] disclosing multi task and mixture of experts, paragraphs[0036]-[0042] disclosing training plurality of experts).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the MoE, as taught by Li, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including ability for large volumes to be handled and optimized ranking objectives where parameters are automatically learned to share across potentially conflicting objectives and bring improvements to full recommendation system. See at least paragraph [0026], [0030], [0069].
Regarding Claims 3, 10, and 17, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 2, 9, and 16. Additionally, Neither Javari nor Ulrich expressly provide for wherein the predicting the next item to be recommended to the user comprises: receiving the embeddings as input; generating modified embeddings based on the embeddings in accordance with knowledge learned during training on how different types of information interact; routing the modified embeddings to at least some of the plurality of experts according to routing knowledge learned in training; and processing recommendations from the at least some of the plurality of experts to determine the next item to be recommended to the user.
However, Li discloses wherein the predicting the next item to be recommended to the user comprises: receiving the embeddings as input; generating modified embeddings based on the embeddings in accordance with knowledge learned during training on how different types of information interact; routing the modified embeddings to at least some of the plurality of experts according to routing knowledge learned in training; and processing recommendations from the at least some of the plurality of experts to determine the next item to be recommended to the user (See at least paragraph [0007] disclosing converting features form one or more data sources into hidden embeddings, MMOE layer comprising multiple experts, generating by each of the multiple experts a prediction input, [0030] disclosing multiple actions learned with multigate mixture of experts for more efficient learning curve to be formed while performances keep improving as well, [0036], [0041] disclosing state are used as input for each expert while embeddings corresponding to actions may be used for labels for training etc., [0042] disclosing each action is regarded as a task for an expert and during training for the expert, the loss calculation for each action is made independently and feature embedding may be shared across the actions to control the interactions between actions).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the experts knowledge learned during training, as taught by Li, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including ability for large volumes to be handled and optimized ranking objectives where parameters are automatically learned to share across potentially conflicting objectives and bring improvements to full recommendation system. See at least paragraph [0026], [0030], [0069].
Regarding Claims 4, 11, and 18, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 1, 8, and 15. Additionally, Javari discloses creating a next sequence data associated with the current event, including information about the UI, the current event information, and the next item recommended to the user; incorporating new performance data related to the next item in the next sequence data obtained by monitoring the user’s activity directed to the next item; adding the new sequence data to the historic sequence data to generate updated historic sequence data; and adapting model via training using the updated historic sequence data (See at least paragraph [0046]-[00650] disclosing model may be trained based on user sequences based on user activity, training data include activity through the interface, determine user interactions with the interface such as clicks, purchases, ATC, etc. in given session, and training model by predicting next user action in reposen to each new user action based on the sequence of prior user actions, etc.).
Neither Javari nor Ulrich expressly provides for a mixture of expert (MoE) prediction model. However, Li discloses a mixture of expert (MoE) prediction model (See at least paragraph [0006], [0007], [0026], [0030], [0069]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the MoE, as taught by Li, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including ability for large volumes to be handled and optimized ranking objectives where parameters are automatically learned to share across potentially conflicting objectives and bring improvements to full recommendation system. See at least paragraph [0026], [0030], [0069].
Regarding Claims 5 and 12, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 1 and 8. Additionally, Javari discloses wherein the current event information includes information related to at least one of: searches conducted by the user during the communication session; one or more items the user exhibits interest via the user’s interactions on the UI; and context of the current event (See at least paragraph [0012], [0015], [0017], [0022], [0029], [0049]-[0050] disclosing current events such as clicks, add to carts, purchases, etc. with items in user interface).
Regarding Claims 6 and 13, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 1 and 8. Additionally, Javari discloses wherein the historic sequence data includes a sequence of events, each of which includes a UI, contextual information, an item recommended on the UI given the context, and performance data associated with the recommended item (See at least paragraph [0017], [0025], [0049] disclosing historic information such as context, sequence information, user interface interactions such as ATC, CTR, and whether or not purchases were made of recommended items).
Regarding Claim 19, Javari, Ulrich, and Li teach or suggest all of the limitations of claim 15. Additionally, Javari discloses wherein the current event information includes information related to at least one of: searches conducted by the user during the communication session, one or more items the user exhibits interest via the user’s interactions on the UI, and context of the current event; (See at least paragraph [0012], [0015], [0017], [0022], [0029], [0049]-[0050] disclosing current events such as clicks, add to carts, purchases, etc. with items in user interface) and the historic sequence data includes a sequence of events, each of which includes a UI, contextual information, an item recommended on the UI given the context, and performance data associated with the recommended item (See at least paragraph [0017], [0025], [0049] disclosing historic information such as context, sequence information, user interface interactions such as ATC, CTR, and whether or not purchases were made of recommended items).
Regarding Claims 7, 14, and 20, Javari, Ulrich, and Li teach or suggest all of the limitations of claims 1, 8, and 15. Javari does not expressly provide for wherein each of the tokens in the task sentence includes at least an item and a performance metric to be used to evaluate whether the user is to achieve a corresponding performance if the item is recommended to the user. However, Ulrich discloses wherein each of the tokens in the task sentence includes at least an item and a performance metric to be used to evaluate whether the user is to achieve a corresponding performance if the item is recommended to the user (See at least paragraph [0067], [0070], [0085], [0086] disclosing token sequence generated as multimodal output, [0090], [0091] disclosing user input prompt tokens which represent users desired configuration, etc., [0101], [0105], [0116] disclosing feedback tokens, [0122] disclosing user response and actions may be recorded and combined with tokens, [0123], [0128], [0359] disclosing string of tokens form prompt composer and generating feedback related to the specific task, [0429]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Javari with the tokens, as taught by Ulrich, since such a modification would have only united elements of the prior art, with no change in their respective functions, and the combination would have yielded predictable results including better prompts for the model based on information other than just natural language provided by a user. See at least paragraph [0005], [0085].
Conclusion
The references cited in the form PTO-892 were not applied under relevant section §103 in the above Office Action, however, they are considered relevant to both claimed and unclaimed features of the instant invention. Applicant is herein advised to review the cited prior art references prior to responding to the instant Office Action in order to expedite prosecution of the instant application. For example:
Billman et al. (US 2025/0147954) disclosing receiving current event information, receiving past history information, and creating a sentence with a number of tokens to represent an embedding space.
Rath et al. (US 2024/0037631) disclosing recommendation system that receives sequence of items selected by user and uses that to help recommend further items for purchase.
“Recommendation for New Users and New Items via Randomized Training and Mixture-of-Experts Transformation” (Zhu, Z., Sefati, S., Saadatpanah, P. and Caverlee, J., Recommendation for new users and new items via randomized training and mixture-of-experts transformation, 2020, July, In Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval (pp. 1121-1130).) disclosing a MoE transformation mechanism to provide personalized item recommendations.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRITTANY E BARGEON whose telephone number is (571)272-2861. The examiner can normally be reached Monday-Friday 9:00am to 6:00pm.
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, Jeffrey A Smith can be reached at (571) 272-6763. 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.
/B.E.B/Examiner, Art Unit 3688
/Jeffrey A. Smith/Supervisory Patent Examiner, Art Unit 3688
Prosecution Insights