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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 5/23/2024 was filed in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
1: Claim(s) 1, 2, 9-13 and 18 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by CN 117272130 A Tan et al (for examination purposes the examiner has relied upon the provided English translation).
2: As for Claim 1, Claim 1 is rejected for reasons discussed related to Claim 13.
3: As for Claim 2, Tan et al teaches on Pages 5, 6 and 13 wherein to generate the de-biased versions of the relevance scores (the output of the main tower network and the debias tower network obtains the prediction result through the prediction network after the polymerization), the isotonic layer simulates a step function. Tan et al teaches re-training the model according to the gating vector weight obtained in the searching stage; in the re-training stage, discarding the characteristic that the gating weight is 0, only reserving the characteristic that the gating vector is greater than the corresponding position of 0 (viewed as a step function); in the re-training stage, stopping updating the weight of the gating vector and only selecting the characteristic.
4: As for Claim 9, Tan teaches on Page 4, wherein the isotonic layer of the deep learning model connects an output layer of the score prediction tower with an output layer of the bias prediction tower. Tan teaches a model re-training stage: inputting the user commodity characteristic into the main tower (score prediction tower), inputting the combined characteristic into the debiasing tower, acting the mask vector obtained in the model searching training stage on each combined characteristic, selecting the characteristic more useful for the model training until the whole model is converged (data from the two towers are converged.
5: As for Claim 10, Tan teaches on Page 4 ”the invention provides a recommendation system click prediction method based on feature selection depolarization, the method is based on the classical double tower model, the bias feature is independently modelled, by adding the combined feature with important deviation information and the method of feature selection in the training process, The deviation removing method is applied to the deep recommendation model, the deviation information is effectively learned in the training process of the model, the influence of the data deviation to the recommendation result is reduced, and more accurate CTR prediction is realized.”. Therefore, Tan teaches wherein the score prediction tower generates the predicted relevance scores independently of the bias prediction tower of the deep learning model.
6: As for Claim 11, Tan teaches on Page 3 the invention provides a recommendation system click prediction method based on feature selection depolarization, the method is based on the classical double tower model, the bias feature is independently modelled, by adding the combined feature with important deviation information and the method of feature selection in the training process. Therefore, Tan teaches wherein the bias prediction tower generates the bias prediction embeddings independently of the scoring tower.
7: As for Claim 12, Tan teaches on Pages 2 and 3 the deep recommendation model often uses double-tower model for training. Tan further teaches providing the de-biased versions of the relevance scores for use by a presentation mechanism (the system updates the model recommendations based on the debias tower) to configure the presentation of the content items (items to be clicked by the user) with a plurality of bias-inducing elements (items recommended) in accordance with the de-biased versions of the relevance scores. The recommendation system updates the items to be clicked based on the combination of the data from the two towers.
8: As for Claim 13, Tan teaches on Pages 3 and 7 and a neural network that performs processing and has stored data. Therefore, Tan teaches A system comprising: at least one processor; and at least one memory coupled to the at least one processor, wherein the at least one memory comprises at least one instruction that, when executed by the at least one processor, cause the at least one processor to perform at least one operation. Tan teaches on Page 2 using a score prediction tower of a deep learning model to generate and output predicted relevance scores for respective content items (tan teaches preference score of the user to the commodity, the preference score through sigmoid activation function to obtain the probability that the user clicks the commodity); Tan teaches in Page 4 using a bias prediction tower of the deep learning model to generate and output bias prediction embeddings (tan teaches a model re-training stage: inputting the user commodity characteristic into the main tower, inputting the combined characteristic into the debiasing tower); Tan teaches in Page 12 using an isotonic layer of the deep learning model to combine the relevance scores output by the score prediction tower with the bias prediction embeddings output by the score prediction tower (tan teaches inputting the combined characteristic to the characteristic importance evaluation model for training, evaluating the deviation importance of each group of characteristic through the score of the combined characteristic by the model); by the isotonic layer (Tan teaches on Page 2 a multi-layer perception network), generating and outputting de-biased versions of the relevance scores based on the combination of the relevance scores with the bias prediction embeddings (tan teaches on Page 4 a model re-training stage: inputting the user commodity characteristic into the main tower, inputting the combined characteristic into the debiasing tower, acting the mask vector obtained in the model searching training stage on each combined characteristic, selecting the characteristic more useful for the model training until the whole model is converged.); and providing the de-biased versions of the relevance scores for use by at least one application, system, model, service, process, or device (the updated model is used by the system).
9: As for Claim 18, Claim 18 is rejected for reasons discussed related to Claim 13.
Allowable Subject Matter
Claims 3-8, 14, 15-17, 19 and 20 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
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JAMES M HANNETT whose telephone number is (571)272-7309. The examiner can normally be reached 8:00 AM-5:00 PM Monday thru Thursday.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Twyler Haskins can be reached at 571-272-7406 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).
/JAMES M HANNETT/Primary Examiner, Art Unit 2639
JMH
August 10, 2026
Prosecution Insights