DETAILED ACTION
This is in reference to communication received 13 April 2026. Claims 1 – 20 are pending for examination. 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 (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Independent claim 11, representative of claim 1, in part is directed toward a statutory category of invention, the claim appears to be directed toward a judicial exception namely an abstract idea. Claim 1 recites invention directed to
Obtaining collected historical data, a sampling strategy is selected to obtain a subset of data that are not labeled (e.g., unlabeled data). Obtained subset of unlabeled data is sent to an oracle (e.g., a human) for getting the unlabeled data associated with a label. Labeled data received from the oracle are applied to a machine learning model to adjust parameters of the machine learning models thereby updating the plurality of dynamic weightings of the machine learning model. These limitations describe as marketing/sales/advertising activities. Obtaining collected data (such as from a file in a filing cabinet), identifying unlabeled data records from the obtained data, and forwarding to an oracle (e.g., a human) asking them to assigns labels to the unlabeled data, which is then used to update the machine learning model.
In addition, using a supervised machine learning methodology to train a model with subset of data labeled by a human, and then using the trained model to label all of the unlabeled data. Not only do these features fail to integrate the abstract idea into a practical application (see below), but it can also reasonably be seen as the conventional application of well-known machine learning concepts to build and train a model to implement the abstract idea on a computer, and merely uses a computer as a tool to perform the abstract idea. See MPEP 2106.05(f).
Represented claim 1, which do recite statutory categories (machine, product of manufacture, for example), the same analysis as above applies to these claims since the method steps are the same. However, the judicial exception is not integrated into a practical application. These claims add the generic computer components (additional elements) of a system comprising one or more hardware processors and a memory (claim 1) to perform the method addressed above (claim 11).
As for dependent claims 2 – 9 and 12 - 19, these claims recite limitations that further define the same abstract idea of using different weighting algorithms for determination of preferences, defining what information will be predicted using the model, defining criteria for rate-eligibility, defining which users will be eligible for discounted-rate, defining filtering criteria for filtering client who should not be considered to offer discounted-rate, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of organizing certain methods of human activity related to advertising, marketing or sales activities or behaviors but for the recitation of generic computer components. Accordingly, the claim recites an abstract idea.
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 – 2, 5 and 11 – 12 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Duesterwald US Publication 2022/0180230 and Tara Wildt published article “What is Human-Guided Machine Learning?” hereinafter referred to as “Wildt”.
Regarding claim 1 and representative claim 11, Duesterwald teaches machine learning based techniques for decision making tasks such as data labeling using a hybrid approach where cooperation between an artificial intelligence (AI) assessor and a human labeler controls automation of the process comprising:
a communication interface [Duesterwald, 0086];
a memory storing instructions [Duesterwald, 0086, claim 18];
one or more processors coupled to the communications interface and to the memory, the one or more processors configured to execute the instructions to perform operations to train a prediction model using active learning [Duesterwald, 0084] comprising:
obtaining, from a database, a set of unlabeled data comprising a set of first data entries for a plurality of clients relating to eligibility for discounted rates (Duesterwald, Namely, in step 604, the AI label predictor selects a batch of unlabeled data items and presents the selected unlabeled data items to the human labeler.) [Duesterwald, 0069];
selecting, based on a plurality of dynamic weightings corresponding to a plurality of sampling strategies, at least one of the plurality of sampling strategies (Duesterwald, the AI label predictor uses a selection strategy such as low margin active learning (see above) to select the unlabeled data items to present to the human labeler.) [Duesterwald, 0069];
obtaining, based on the at least one of the plurality of sampling strategies, a subset of the set of unlabeled data (Duesterwald, Namely, in step 604, the AI label predictor selects a batch of unlabeled data items and presents the selected unlabeled data items to the human labeler.) [Duesterwald, 0069];
sending, to the communication interface, the subset of unlabeled data for investigation by the oracle (Duesterwald, the AI assistant selects a batch of the (unprocessed) input data items and presents the data items to the human decision maker) [Duesterwald, 0036, 0038];
receiving, from the communication interface, a set of labeled data comprising second data entries corresponding to the subset of unlabeled data, each of the second data entries comprising a label: from the oracle for a corresponding first data entry from the subset of unlabeled data (Duesterwald, In step 206, (true) human decisions for the data items presented to the human decision maker are obtained from the human decision maker. These human decisions for the data items are then used by the AI assistant in making further decision predictions, and so on) [Duesterwald, 0040];
applying, in a training phase of the machine learning model, the set of labeled data to the machine learning model to adjust parameters of the machine learning model (Duesterwald, Based on the presented analysis from the AI assessor (which is provided at the end of each active learning loop iteration), if it is decided in step 208 that NO automation readiness has not yet been achieved, then delegation to the AI assistant is declined and training is continued with manual decision in the next iteration of the active learning refinement loop.) [Duesterwald, 0044]; and
Duesterwald does not explicitly teach updating on weightings. However, Wildt teaches there are three general types of machine learning: supervised, unsupervised, and reinforcement learning. Human-guided machine learning is a type of supervised learning, which uses a set of human-labeled training data to develop a model. …. Human-guided machine learning is a process whereby subject matter experts accelerate the learning process by teaching the technology in real-time [Wildt, page 2].
Therefore, at the time of filing, it would have been obvious to one of ordinary skill in the art to modify Duesterwald by adopting teachings of Wildt to reduce amount of time humans need to spend performing a specific task as the machine accuracy increases.
Duesterwald in view of Wildt teaches system and method further comprising:
updating, based at least on the set of labeled data, the plurality of dynamic weightings (Wildt, if the machine learning model comes across a piece of data it is uncertain about, a human can be asked to weigh in and give feedback. The model then learns from this input, and uses it to make a more accurate prediction the next time. Human-guided machine learning works from the bottom up by first using algorithms to conduct the heavy lifting of identifying relationships within the data, and engaging humans when necessary for training or validation.) [Wildt, page 2].
Regarding claim 2 and representative claim 12, as combined and under the same rationale as above, Duesterwald and Wildt teaches system and method, wherein the selection model comprises an exploration weighting, wherein the plurality of dynamic weightings includes an exploration weighting, the exploration weighting biasing the selection model to selecting unlabeled data that the machine learning model is not trained for (Wildt, if the machine learning model comes across a piece of data it is uncertain about, a human can be asked to weigh in and give feedback. The model then learns from this input, and uses it to make a more accurate prediction the next time. Human-guided machine learning works from the bottom up by first using algorithms to conduct the heavy lifting of identifying relationships within the data, and engaging humans when necessary for training or validation.) [Wildt, page 2].
Regarding claim 5 and representative claim 15, Duesterwald and Wildt teaches system and method, wherein the operations further comprise applying the prediction model to the set of unlabeled data to a set of predictions, wherein the set of predictions comprises a prediction of discounted rate eligibility for each of the plurality of clients (Duesterwald, On the other hand, if it is decided in step 208 that YES automation readiness has been achieved, then the decision making capabilities are fully delegated to the AI assistant and in step 210 the AI assistant performs the decision making on the remainder of the data tasks without receiving verification (agreed/not agreed) from the human decision maker.) [Duesterwald, 44].
Claims 3 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Duesterwald US Publication 2022/0180230 and Tara Wildt published article “What is Human-Guided Machine Learning?” hereinafter referred to as “Wildt” and Julian Bright published article “Dynamic A/B testing for machine learning models with Amazon SageMaker MLOps projects” hereinafter referred to as “Bright”.
Regarding claim 3 and representative claim 13, Duesterwald in view of Wildt does not teach dynamic weightings to include challenger weighting. However, Bright teaches Before deploying this model to all users, it’s a good idea to run this new or “challenger” model side-by-side with an existing “champion” model in an A/B test to find empirical evidence of the impact this new model has on your business metrics. …. When you’re confident that this new challenger model is the outperforming your previous model, you can deploy this new model to all users, and begin the process again.
Therefore, at the time of filing, it would have been obvious to one of ordinary skill in the art to modify Duesterwald in view of Wildt by adopting teachings of Bright to in deploy new model with higher confidence of achieving desired results.
as combined and under the same rationale as above, Duesterwald in view of Wildt and Bright teaches system and method, wherein the plurality of dynamic weightings includes a challenger weighting, the challenger weighting biasing the selection model to randomly selecting unlabeled data (Bright, Before deploying this model to all users, it’s a good idea to run this new or “challenger” model side-by-side with an existing “champion” model in an A/B test to find empirical evidence of the impact this new model has on your business metrics, such as click-through rate, conversion rate, or revenue. By collecting real-time feedback as your model is running, you can optimize how traffic is distributed between the champion and challenger models of the period of the test, which can often run for several weeks. When you’re confident that this new challenger model is the outperforming your previous model, you can deploy this new model to all users) [Bright, page 2].
Claims 4, 6, 14 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Duesterwald US Publication 2022/0180230 and Tara Wildt published article “What is Human-Guided Machine Learning?” hereinafter referred to as “Wildt” and Underfitted YouTube video “Active Learning – The Secret of Training Models Without Labels” hereinafter referred to as “Underfitted”.
Regarding claim 4 and representative claim 14, Duesterwald in view of Wildt does not explicitly teach dynamic weightings to include exploitation weighting However, Underfitted teaches several classic exploration algorithms that work out pretty well in the multi-armed bandit problem or simple tabular RL to find the best solution as fast as possible [Underfitted, page 1].
Therefore, at the time of filing, it would have been obvious to one of ordinary skill in the art to modify Gong in view of Duesterwald by adopting teachings of Underfitted to find the best solution as fast as possible [Underfitted, page 1].
as combined and under the same rationale as above, Duesterwald and Wildt and Underfitted teaches system and method, wherein the plurality of dynamic weightings includes an exploitation weighting corresponding to a particular one of the plurality of sampling strategies, the particular one of the sampling strategies relating to selecting unlabeled data based on exploitation sampling (Underfitted, Exploitation and Exploration strategies are used in Deep Reinforcement Learning) [Underfitted, page 1].
Regarding claim 6 and representative claim 16, as combined and under the same rationale as above, Duesterwald and Wildt and Underfitted teaches system and method, wherein the operations further comprise providing a reward score of the machine learning model, the reward score being based on the set of predictions (Underfitted teaches providing score) [Underfitted, page 20], wherein the reward score is a number of clients in the plurality of clients that require a re-rate based on an associated discounted rate eligibility (Underfitted, Count-based Exploration) [Underfitted, page 3, 4].
Claims 7 – 10 and 17 – 20 are rejected under 35 U.S.C. 103 as being unpatentable over Duesterwald US Publication 2022/0180230 and Tara Wildt published article “What is Human-Guided Machine Learning?” hereinafter referred to as “Wildt” and Singh US Publication 2020/0320565
Regarding claim 7 and representative claim 17, Duesterwald and Wildt does not teach using unlabeled data for renewals within a renewal period. However, Singh teaches system and method to determine a data reward to provide to a mobile device customer in response to the occurrence of the event associated with the renewal of a current prepaid service plan.
Therefore, at the time of filing, it would have been obvious to one of ordinary skill in the art to modify Duesterwald and Wildt by adopting teachings of Singh put in use their machine-learning algorithm in services industries to help them identify customer who are up for renewal before their subscription expires.
as combined and under the same rationale as above, Duesterwald and Wildt and Singh teaches system and method, wherein the set of unlabeled data is for renewals within a renewal period (Singh teaches system and method to determine a data reward to provide to a mobile device customer in response to the occurrence of the event associated with the renewal of a current prepaid service plan) [Singh, 0014].
Regarding claim 8 and representative claim 18, as combined and under the same rationale as above, Duesterwald and Wildt and Singh teaches system and method, wherein the set of unlabeled data includes data relating to clients and the operations further comprise removing the first data entries in the set of unlabeled data for the clients over a threshold age (Singh, In order to identify customers eligible or otherwise targeted for renewal rewards, the customer incentive system 150 may access and/or receive information from the customer data system 140 and/or the database 160, such as information that identifies the customer, information that identifies the customer's activities within the network 120, information that tracks historical preferences or attributes assigned to the customer, information that groups the customer into various groups or cohorts of customers of the carrier, and so on. …. Reward determination module 220 may follow certain parameters or instructions when determining customer eligibility for rewards.) [Singh, 0025, 0035].
Regarding claim 9 and representative claim 19, as combined and under the same rationale as above, Duesterwald and Wildt and Singh teaches system and method, wherein the set of unlabeled data includes data relating to clients and the method further comprises receiving an affinity list from the database and searching the first data entries for the clients in the affinity list, wherein any of the first data entries for the clients in the affinity list are removed from the set of unlabeled data (Singh, In order to identify customers eligible or otherwise targeted for renewal rewards, the customer incentive system 150 may access and/or receive information from the customer data system 140 and/or the database 160, such as information that identifies the customer, information that identifies the customer's activities within the network 120, information that tracks historical preferences or attributes assigned to the customer, information that groups the customer into various groups or cohorts of customers of the carrier, and so on.) [Singh, 0025].
Regarding claim 10 and representative claim 20, as combined and under the same rationale as above, Duesterwald and Wildt and Singh teaches system and method, wherein the set of unlabeled data includes data relating to clients and the method further comprises receiving a member list from the database and searching the first data entries for the clients in the member list, wherein any of the first data entries for the clients in the member list are removed from the set of unlabeled data (Singh, In order to identify customers eligible or otherwise targeted for renewal rewards, the customer incentive system 150 may access and/or receive information from the customer data system 140 and/or the database 160, such as information that identifies the customer, information that identifies the customer's activities within the network 120, information that tracks historical preferences or attributes assigned to the customer, information that groups the customer into various groups or cohorts of customers of the carrier, and so on.) [Singh, 0025].
Response to Arguments
Applicant's arguments filed 13 August 2026 have been fully considered, however, applicant’s arguments are for amended claimed invention. While performing an updated search in view of amended claimed invention, a new prior art was found and cited in this office action. Therefore, applicant’s arguments are moot under new grounds of rejection.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Naresh Vig whose telephone number is (571)272-6810. The examiner can normally be reached Mon-Fri 06:30a - 04:00p.
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, Ilana Spar can be reached at 571.270.7537. 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.
/NARESH VIG/Primary Examiner, Art Unit 3622
June 16, 2026
Prosecution Insights