DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is responsive to communications: Amendment filed on 11/10/2025.
Claims 1, 3, 5, 7, 9, 11, 13, 15, 17, and 27-30 are pending. Claims 1, 7, 13, are independent. Claims 4, 10, 16, 19-26 are previously canceled. Claims 2, 6, 8, 12, 14, and 18 are newly canceled.
The previous rejection of claims 1, 3, 5, 7, 9, 11, 13, 15, 17, and 27-30 under 35 USC § 103 have been withdrawn in view of the amendment.
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.
Claim(s) 1, 3, 5, 7, 9, 11, 13, 15, 17, 27, and 29-30 is/are rejected under 35 U.S.C. 103 as being unpatentable over Klenner et al. (US2018/0357343) in view of Ruben et al. (“Hybrid data-driven physics model-based framework for enhanced cyber-physical smart grid security”) and Xin et al. (US2019/0121336).
In regards to claim 1, Klenner et al. substantially discloses a computer-implemented method for balancing use of a surrogate model with a need for accuracy in forecasting, the method comprising:
Receiving, by a computing system comprising a deep learning based surrogate model and a physics based mathematical model, input boundary conditions to predict future conditions of a complex weather problem (Klenner et al. fig 3 302, para[0055], receives one or more data samples, para[0061] examples of complex problems include predicting future production);
Determining, by the comparison module, based on the resulting similarity, whether the [neural network] based surrogate model is reliable for simulating the future condition at the particular time window of the complex weather problem (Klenner et al. para[0053] ln6-11, determines if data is within region of competence for data-driven model);
Outputting, by the computing system, the second set of results as the system output for simulating the future condition at the particular time window of the complex problem (Klenner et al. para[0054] ln7-15, runs a physics-based simulation if data driven model competence level is below threshold);
Klenner et al. does not explicitly disclose generating, by the deep learning based surrogate model, a first set of results simulating the future conditions of the complex weather problem based on the input boundary conditions, the future conditions representing a first period of time including a particular time window of the complex weather problem;
generating, by the physics based mathematical model, a second set of results simulating a future condition of the particular time window based on the input boundary conditions, the particular time window representing a partially run analysis for a duration shorter than the first period of time;
receiving, by a comparison module of the computing system, the first set of results and the second set of results for evaluation at the particular time window;
Responsive to receiving the sets of results, quantifying, by the comparison module, a similarity of the first set of results to the second set of results using an L^2-norm or a Mahalanobis distance to asses reliability of the deep learning based surrogate model for simulating additional future conditions in other complex weather problems;
switching, by the comparison module, outputs for the particular time window from the deep learning based surrogate model to the physics based mathematical model, based on the resulting similarity being below a given threshold.
However Ruben et al. discloses generating, by the deep learning based surrogate model, a first set of results simulating the future conditions of the complex weather problem based on the input boundary conditions, the future conditions representing a first period of time including a particular time window of the complex weather problem (Ruben et al. pg4 section 3.2 para1, in addition to spatial information that is used in physics based model, data-driven machine learning algorithms are proposed to take advantage of both temporal and spatial information);
generating, by the physics based mathematical model, a second set of results simulating a future condition of the particular time window based on the input boundary conditions, the particular time window representing a partially run analysis for a duration shorter than the first period of time (Ruben et al. pg4 section 3.1 para 1, In this approach, the system is modeled as a set of non-linear algebraic equations based on the physics of the system);
receiving, by a comparison module of the computing system, the first set of results and the second set of results for evaluation at the particular time window (Ruben et al. fig. 1 pg6 section3.3 para1, in order to employ the anomaly detection capabilities of both state estimator);
Responsive to receiving the sets of results, quantifying, by the comparison module, a similarity of the first set of results to the second set of results using an L^2-norm or a Mahalanobis distance to asses reliability of the deep learning based surrogate model for simulating additional future conditions in other complex weather problems (Ruben et al. pg6 section 3.3 para2, To find the overall ECD decision score for each testing sample, we consider the squared Mahalanobis distance of Local CorrDet detector which led us to decide whether the sample was anomalous or normal);
switching, by the comparison module, outputs for the particular time window from the deep learning based surrogate model to the physics based mathematical model, based on the resulting similarity being below a given threshold (Ruben et al. pg6 section4.1 para1, Error detection based on physics-based models is performed as a post processing step to PSSE using CMECST).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the hybrid detection model of Ruben et al. in order to explore both the temporal and spatial characteristics of the model environment (Ruben et al. pg4 section 1 para9).
Klenner et al. does not explicitly disclose in response to determining that the deep learning based surrogate model is not reliable, stopping, by the computing system, running of the deep learning based surrogate model and triggering, by the computing system, running of the physics based mathematical model for simulating the additional future conditions of the other complex weather problems; and
In parallel with the triggering of running the physics based mathematical model for simulating the additional future conditions, training and updating, by an online training module of the computing system, the deep learning based surrogate model using the second set of the results generated by the physics based mathematical model.
However Xin et al. substantially discloses in response to determining that the deep learning based surrogate model is not reliable, stopping, by the computing system, running of the deep learning based surrogate model and triggering, by the computing system, running of the physics based mathematical model for simulating the additional future conditions of the other complex weather problems (Xin et al. para[0022]-[0023], retrieves operational data from physical model of a peer power plant identified as the same type as the new power plant); and
In parallel with the triggering of running the physics based mathematical model for simulating the additional future conditions, training and updating, by an online training module of the computing system, the deep learning based surrogate model using the second set of the results generated by the physics based mathematical model (Xin et al. para[0033], error score is used to update source data used to train model. Prediction model may be loaded with source data 310, source data 310 may comprise data from a peer physics model).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the heterogenous modeling of Xin et al. in order to accurately model systems with limited historical data (Xin et al. para[0003]).
In regards to claim 3, Klenner et al. as modified by Ruben et al. and Xin et al. substantially discloses the computer-implemented method of claim 1, wherein the deep learning based surrogate model is trained offline before deployed as a default choice for simulating the future conditions of the complex problem, wherein the deep learning based surrogate model is trained offline with historical data from the physics based mathematical model (Klenner et al. para[0051], historical data collected from sensors associated with asset may be used to train the model).
In regards to claim 5, Klenner et al. as modified by Ruben et al. and Xin et al. substantially discloses the computer-implemented method of claim 1, further comprising: in response to determining that the deep learning based surrogate model is reliable:
continuing to run the deep learning based surrogate model as a default choice for simulating the future conditions of the complex problem and stopping partially running the physics based mathematical model (Klenner et al. para[0057], determines data-driven model is sufficient to facilitate analysis of a physical phenomenon); and
Outputting the results of running the deep learning based surrogate model as system outputs of simulating the future conditions of the complex problem (Klenner et al. fig. 3 para[0064], the generated hybrid output may be transmitted to user display).
In regards to claims 7, 9, 11, claim 7 recites substantially similar limitations to claim 1, claim 9 recites substantially similar limitations to claim 3, claim 11 recites substantially similar limitations to claim 5. Thus claims 7, 9, and 11 are rejected along the same rationale as claims 1, 3 and 5.
In regards to claims 13, 15, and 17, claim 13 recites substantially similar limitations to claim 1, claim 15 recites substantially similar limitations to claim 3, claim 17 recites substantially similar limitations to claim 5. Thus claims 13-15 and 17-18 are rejected along the same rationale as claims 1-3, and 5-6.
In regards to claim 27, Klenner et al. as modified by Ruben et al. and Xin et al. substantially discloses the computer-implemented method of claim 1, wherein partially running the physics based surrogate model further includes:
Running the model partially in time (Ruben et al. pg4 section 3.1 para3, at each iteration generates a solution for interval).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the hybrid detection model of Ruben et al. in order to explore both the temporal and spatial characteristics of the model environment (Ruben et al. pg4 section 1 para9).
In regards to claim 29, Klenner et al. as modified by Ruben et al. and Xin et al. discloses the computer implemented method of claim 1, wherein quantifying the similarity includes computing a performance score indicating a degree of similarity (Ruben et al. pg6 section 3.3 para4, generates fusion score from comparison of state estimator solution and CorrDet algorithm).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the hybrid detection model of Ruben et al. in order to explore both the temporal and spatial characteristics of the model environment (Ruben et al. pg4 section 1 para9).
In regards to claim 30, Klenner et al. as modified by Ruben et al. and Xin et al. discloses the computer implemented method of claim 1, further comprising:
Receiving, periodically, additional input conditions for an additional complex problems (Ruben et al. pg6 section 4.1 para1, the mean loading condition is updated periodically to model physical reality as aptly as possible);
Generating, by the physics based mathematical model, a third set of results simulating another future condition at another time window of the complex problem (Ruben et al. pg7 section 4.2 para3, score for Ensemble CorrDet detector is better than score for state estimator);
Determining, by the comparison model, based on a comparison of the third set of results to a simulation by the deep learning based surrogate model, whether the deep learning based surrogate model has diverged from the physics based mathematical model (Ruben et al. pg7 section 4.2 para3, the proposed hybrid data-driven physics model-based anomaly detection improved score compared to state estimator results).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the hybrid detection model of Ruben et al. in order to explore both the temporal and spatial characteristics of the model environment (Ruben et al. pg4 section 1 para9).
Claim(s) 28 is/are rejected under 35 U.S.C. 103 as being unpatentable over Klenner et al. in view of Ruben et al., Xin et al. and Sundararaj et al. (US2020/0158004).
In regards to claim 28, Klenner et al. as modified by Ruben et al. and Xin et al. discloses the computer-implemented method of claim 1. Klenner et al. does not explicitly disclose wherein: the complex problem is weather forecasting;
The future conditions are ten days of forecasts; and
The future condition at a particular time window is a one-day forecast.
However Sundararaj et al. discloses wherein: the complex problem is weather forecasting (Sundararaj et al. para[0052], generate weather forecast to determine use of heat exchange) ;
The future conditions are ten days of forecasts (Sundararaj et al. para[0052], generate consecutive 5 day forecasts); and
The future condition at a particular time window is a one-day forecast (Sundararaj et al. para[0052], predict days hot enough to overheat machine).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the optimization method of Klenner et al. with the monitoring method of Sundararaj et al. in order to predict and track the conditions a system is subjected to (Sundararaj et al. para[0009]).
Response to Arguments
Applicant’s arguments, see 9, filed 11/10/2025, with respect to claims 1, 3, have been fully considered and are persuasive. The 101 rejection of 8/13/2025 has been withdrawn.
Applicant’s arguments with respect to claim(s) 1, 3, 5, 7, 9, 11, 13, 15, 17, and 27-30 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 NICHOLAS HASTY whose telephone number is (571)270-7775. The examiner can normally be reached Monday-Friday 8:30am-5:00pm.
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, Matt Ell can be reached at (571)270-3264. 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.
/N.H/Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141