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 .
Response to Arguments
Applicant arguments regarding the 35 USC 101 have been considered but found unpersuasive for the reasons set forth below, infra 35 USC 101 analysis. In particular, a human mind given multiple models may combine such models into an aggregate model, a summation, or a visual representation of hierarchical models while observing model outputs, identifying model contributions, associated production tasks per model, and determining deviations correspond to global changes.
In contrast to Example 39, the Examiner suggests Applicant incorporate an equivalent training limitations such that “claim limitations that only encompass AI in a way that cannot be practically performed in the human mind…” such as involving training the machine learning models for the specific global detection vs the local detection. By specifically linking the training method (e.g. training a neural network is not a mental process, as per example 39), for learning models for determining global vs local deviations of specific production lines, this is not considered a mental process. Moreover, the specific training steps described above would appear to practically apply the abstract idea.
Regarding a technical solution, in light of Applicant’s arguments, pages 14-16, the claim limitations do not distinguish between local and global changes of the production line as discussed in the arguments as well as the determination of time periods when a less efficient model is running relative an updated model. Moreover, the Applicant describes examples of a technical problem including a model update process as discontinuous and sub-optimal models running linger if humans controlled an update process. The specification describes the improvements in the context of time while the claim limitations generally reflect an update occurs in response to a deviation without reflecting a time improvement. The claims recite a comparison framework to determine a deviation, the deviation corresponds to a global change based on metadata, triggering a model update, searching for models, generating an ensemble model, observing model outputs, and adjusting model weights. The combination of limitations reflects a model update process based on deviations but without relating this the improvements described in the specification. The claims generally provide an inference time may be saved through automation but without specifying a basis as to how time is saved. While the claims do not have to expressly recite the improvement, here, the combination of limitations show more of model adjustments for deployment based on a deviation. The claims do not appear to capture the methods within the specification as for providing that time is optimized. For example, it would appear time can be optimized based on “whether fellow screw driving lines detect a similar increase…” and associated “correlations” because a human may fail to observe granular relationships.
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, 28-29, 37-38, and 56-58 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claim(s) recite(s) detecting a deviation between a first output of a first AI model and a baseline metric, determining that the deviation corresponds to a global change based on metadata (claims 1, 2, 10-11, 19-20, 28-29, 37) ; trigger an update based on the deviation (e.g. as interpreted, identifying a need to update, claims 1,2,10-11, 19, 28, 37); searching for candidate models (claims 1, 4, 11, 13, 37); generate an ensemble model (e.g. as interpreted, grouping multiple models, claims 1, 10, 37); adjusting model weights (claims 1,11, 18, 37, 56); identifying similar models (claims 6, 8, 15, 17); identifying differences (claim 7) (claims 1, 11, 37 ); determining combined predictions (claim 9), and determining a global metric based on first and second ensemble AI models, and generating “closer predictions” (claims 57-58). Each of the limitations involve opinion, observation, and judgment. The determination of model selection is based in part on identifying deviations while also determining a model requires updating. The grouping or averaging models to form an ensemble model represents a mental process of adding and grouping. The updating is a determination result for replacing a prior model involving judgment, MPEP 2106.04(a)(2).
This judicial exception is not integrated into a practical application because the limitations comprising monitoring output, obtaining candidate models, deploying an ensemble model, updating an ensemble model; applying operating data from a production line, collecting metadata (e.g. sensor data); implementing different model architectures, including providing model outputs (claims 1-20, 28-29, 37-38, and 56), and storing (claims 57-58) represent insignificant extra-solution activity. Additionally, executing the first ensemble AI model to generate predictions from the candidate AI models represents an application of the AI models. The inclusion of a processor, circuitry, and memory represent mere instructions to apply the abstract idea. The deployment, selection, and updating are generally recited so as to apply the abstract idea. Each of the production line, sensors, models, and associated data are generally recited so as to generally link the abstract idea to the field of model selection. Moreover, the association of tasks with production lines is recited generally so as to generically link the abstract idea to the field of factory models.
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the inclusion of a processor, memory, and circuitry represent mere instructions to apply the abstract idea while the data monitoring, collection, storing, updating, deployment are well-known, conventional, and routine, see MPEP 2106.05(d), infra applied prior art, references cited: 20210117869 20180060330 20220122000 20220012641 20220122000 20230148321 20220391760 20210304151 (e.g. model scoring, grouping, averaging, updating, selection, and deployment).
Moreover, triggering an automated model update as well as executing and updating ensemble models represent insignificant extra-solution considered well-known, conventional, or routine, see 14/659814, 16/812105 (e.g. see adjusting model weights)
It is noted, in light of Example 39, “training a neural network for facial detection,” the claim does not recite any of the judicial exceptions enumerated in the 2019 PEG. For instance, the claim does not recite any mathematical relationships, formulas, or calculations. While some of the limitations may be based on mathematical concepts, the mathematical concepts are not recited in the claims. Further, the claim does not recite a mental process because the steps are not practically performed in the human mind. Finally, the claim does not recite any method of organizing human activity such as a fundamental economic concept or managing interactions between people. Thus, the claim is eligible because it does not recite a judicial exception.”
The Examiner suggests incorporating a training limitation to preclude the application of a mental process.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
See general model updating, accuracy assessment, drift analysis, assembly line models, maintaining model repositories, hyperparameter analysis
20210390455 20210357785 20090106178 20230080357 20240176732 20230186175 20220230024 11315039 20220067573 20210201209 20200387836 20200151619 20190102361 20190102700 10209974 20120284212 20120284213 20200387836 11227188 11544494 20210201209 20240104394 20220222927 20210325861 20210125066 20220067622 20210201209 20240185576 20220147672 20240185117 20240061410 (lines with shared model) 20230185652 20230097885 20220277231 ( model update) 20220269248 20220138574 11055639 10783469 20190094843 11640447 20210358065 20200125900 20220138574 20240061410 20220067622 20220051140 -0037-38 0049 20200125900 -0029 20190155717 -0066 10089383 20220199266 20220199266 20220076164 20220067428- 20220318689 2022051140
See ensemble model scores and model selection
20210117869 20180060330 20220122000
20220012641 20220122000 20230148321 20220391760 20210304151
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DARRIN D DUNN whose telephone number is (571)270-1645. The examiner can normally be reached M-Sat (10-8) PST.
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, Robert Fennema can be reached at 571-272-2748. 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.
/DARRIN D DUNN/ Patent Examiner, Art Unit 2117