DETAILED ACTION
This action is in response to the claims filed 17 November 2023 for application 18513173 filed 17 November 2023. Currently claims 1-20 are pending.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 2-3, 7-8, 10-13, 17, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Kursun (US 20210049455 A1) in view of Badawy et al. (US 11295241).
Regarding claims 2 and 12, Kursun discloses: A method comprising:
detecting, in a production dataset, a data shift from a training dataset used to train a first machine learning model (“receive input data for analysis and expansion; analyze the input data using a machine learning model to identify an emerging pattern in the input data; extract common data characteristics from the identified emerging pattern in order to determine a data scenario;” claim 1);
providing, to a generative adversarial network, the production dataset to cause the generative adversarial network to generate synthetic data based on the production dataset (“utilize encoding and decoding logic to expand scope of the data scenario; and based on the expanded scope of the data scenario, generate a synthetic data set using a generative adversarial neural network.” Claim 1);
determining that a magnitude of the data shift satisfies a first threshold … (Fig 4 test identified abnormal pattern based on generated synthetic data “pass” or “fail” is interpreted as a magnitude and a threshold);
performing, based on determining that the magnitude of the data shift satisfies the first threshold… , a first training routine comprising training a second machine learning model using a first updated dataset, the first updated dataset comprising the training dataset, the production dataset, and the synthetic data (“The synthetic data storage 326 is configured to store synthetically generated data generated by the system (i.e., via synthetic data engine 324). The synthetic data stored in the synthetic data storage 326 may be used for training a machine learning model or artificial intelligence engine, and may also be combined with historical data or user profile data in order to create synthetic profiles, as further discussed in FIG. 8. The synthetic data storage 326 may include adversarial or extrapolated scenarios or data generated by the systems described herein which may be fed back into machine learning models to train the system. In some embodiments, the system 130 may include an adversarial function configured for providing adversarial learning and modeling to the system by introducing unreliable or erroneous data to the system; a learning or adaptation function for defining system response to data changes or an adaptation rate for implementing changes (i.e., model reconfiguration) within an architecture of the systems described herein; and an alertness function and robustness function for defining an appropriate system reaction, response, or extent of system reaction based on one or more environmental conditions or previous interactions. In some embodiments, various synthetic data may be injected in an outgoing data stream in real-time and over multiple iterations in order to further aid in identifying profiling patterns by analyzing the various responses received in correspondence to the synthetic data.” [0042], Fig 4 step 426); and
replacing the first machine learning model with the second machine learning model in a production environment (Fig 4, claim 1, claim 4, note: the original models can be retrained or new models can be created using the synthetic data, historical and new data).
However, Kursun does not explicitly disclose: and does not satisfy a second threshold; and does not satisfy the second threshold.
Badawy teaches: and does not satisfy a second threshold; and does not satisfy the second threshold (“In some embodiments, there may be an additional threshold or range (again used herein interchangeably) associated with a warning zone such that if the drift measure falls within that warning zone range (or above or below that threshold or range, etc.) a warning action may be taken such as raising a notification or alert to a user associated with the enterprise or the provider of the machine learning artificial intelligence system that a data drift is occurring. As another example, there may one or more additional thresholds or ranges associated with a major drift zone such that if the drift measure falls within that major drift zone range (or above or below that threshold or range, etc.) it can be determined that complete retraining of the machine learning model (e.g., initially trained on the first dataset) should be undertaken (e.g., the data of the first dataset and the second dataset are so different that a complete retraining of the machine learning model is needed).” C5L49-65).
Kursun and Badawy are in the same field of endeavor of detecting shifts in data and retraining/replacing models using synthetic data and are analogous. Kursun discloses an exemplary method of replacing and retraining models depending on data drift being detected with a threshold. Badawy teaches two thresholds for issuing a warning and then causing retraining/replacement. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the known data drift and synthetic data model replacement system disclosed by Kursun with the two known thresholds for data drift as taught by Badawy to provide predictable results of a warning period before a costly retraining is needed.
Regarding claims 3 and 13, Kursun discloses: The method of claim 2, further comprising:
detecting, in a new production dataset, a new data shift from the training dataset used to train the first machine learning model;
determining that a new magnitude of the new data shift satisfies the second threshold;
performing, based on determining that the new magnitude of the new data shift satisfies the second threshold, a second training routine comprising training a new machine learning model using a second updated dataset different from the first updated dataset, the second updated dataset comprising the production dataset and the synthetic data; and
replacing the first machine learning model with the new machine learning model.
(Fig 4 multiple models and iterations can be created)
Regarding claim 7, Kursun does not explicitly disclose, however, Badawy teaches:
The method of claim 2, further comprising:
detecting, in the production dataset, a new data shift from the training dataset used to train the first machine learning model;
determining that a new magnitude of the new data shift does not satisfy the first threshold; and
based on determining that the new magnitude of the new data shift does not satisfy the first threshold, refraining from implementing a training routine (“In some embodiments, there may be an additional threshold or range (again used herein interchangeably) associated with a warning zone such that if the drift measure falls within that warning zone range (or above or below that threshold or range, etc.) a warning action may be taken such as raising a notification or alert to a user associated with the enterprise or the provider of the machine learning artificial intelligence system that a data drift is occurring. As another example, there may one or more additional thresholds or ranges associated with a major drift zone such that if the drift measure falls within that major drift zone range (or above or below that threshold or range, etc.) it can be determined that complete retraining of the machine learning model (e.g., initially trained on the first dataset) should be undertaken (e.g., the data of the first dataset and the second dataset are so different that a complete retraining of the machine learning model is needed).” C5L49-65).
Regarding claims 8 and 17, Kursun discloses: The method of claim 2, wherein performing the first training routine comprises assigning first weights to the production dataset and the synthetic data and assigning second weights to the training dataset, wherein the first weights are heavier than the second weights (“The system receives the streaming data where the data is then analyzed and processed by one or more machine learning models for decisioning purposes. Machine learning models, individually and/or structured as clusters, may be trained based on predetermined training data and/or new data acquired in real-time (i.e., from the data stream), wherein the system learns from the data by dynamically identifying patterns as the information is received and processed. In some embodiments of the present invention, machine learning models may be adaptive, wherein the models may be reconfigured based on different environmental conditions and/or an analysis and evaluation of the individual model performance. The model may be modified by the system by having one or more individual models and/or clusters added, removed, made inactive, or the like. In another example, the system may weight particular the conclusions of particular models and/or model clusters more than others. Population architecture refers to a collection and particular arrangement of active machine learning models and/or clusters of machine learning models that are configured to process information mathematically or computationally to make decisions. Particular models and/or clusters may be weighted by the system to emphasize the impact or contribution of the particular models and/or clusters over others.” [0047])
.
Regarding claims 10 and 19, Kursun discloses: The method of claim 2, further comprising calculating the magnitude of the data shift by comparing one or more predictions generated by the first machine learning model with one or more ground truths derived from the training dataset (Fig 4 abnormal pattern is a result not represented in the first training data set).
Regarding claims 11 and 20, Kursun discloses: The method of claim 2, further comprising calculating the magnitude of the data shift using one or more evaluation metrics associated with the first machine learning model (“In some embodiments, “monitoring” may further comprise analyzing or performing a process on something such as a data source or data stream either passively or in response to an action or change in the data source or data stream. In a specific embodiment, monitoring may comprise analyzing performance of one or more machine learning models or engines using performance metrics associated with one or more of the models.” [0023]).
Claim(s) 9 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Kursun in view of Badawy and further in view Suryanarayan (US 20250045626).
Regarding claims 9 and18, Kursun does not explicitly discloses: The method of claim 2, further comprising calculating the magnitude of the data shift using feature similarity scores based on the training dataset and the production dataset.
Suryanarayan teaches: further comprising calculating the magnitude of the data shift using feature similarity scores based on the training dataset and the production dataset (“In some embodiments, a predicted data drift is indicative of a predicted performance of a machine learning model. For instance, a predicted data drift that is indicative of a low rate of dissimilarities (e.g., 90% similarity score, etc.) between the historical training dataset and the contemporary input dataset 312 may be predictive of a high performance (accuracy, etc.) of a machine learning model, whereas a predicted data drift indicative of a high rate of dissimilarities (e.g., 40% similarity score, etc.) between the historical training dataset and the contemporary input dataset 312 may be predictive of a low performance (e.g., inaccuracy, etc.) of a machine learning model.” [0093]).
Kursun, Badawy and Suryanarayan are in the same field of endeavor of detecting shifts in data and retraining/replacing models and are analogous. Kursun discloses an exemplary method of replacing and retraining models depending on data drift being detected with a threshold. Badawy teaches two thresholds for issuing a warning and then causing retraining/replacement. Suryanarayan teaches similarity scores for detecting data drift. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the known data drift and synthetic data model replacement system disclosed by Kursun and Badawy with the known similarity score as taught by Suryanarayan to yield predictable results.
Allowable Subject Matter
Claim 1 is allowed.
Claims 4-6 and 14-16 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. Subhash Dhok et al. (US 20240144662 A1) and Walters et al. (US 20200012900 A1) disclose data drift detection and model retraining.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ERIC NILSSON whose telephone number is (571)272-5246. The examiner can normally be reached M-F: 7-3.
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, James Trujillo can be reached at (571)-272-3677. 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.
/ERIC NILSSON/Primary Examiner, Art Unit 2151
Prosecution Insights