Prosecution Insights
Last updated: April 19, 2026
Application No. 17/948,620

MACHINE LEARNING-BASED DECISION FRAMEWORK FOR PHYSICAL SYSTEMS

Non-Final OA §103
Filed
Sep 20, 2022
Examiner
TANG, MICHAEL XUEFEI
Art Unit
2115
Tech Center
2100 — Computer Architecture & Software
Assignee
International Business Machines Corporation
OA Round
3 (Non-Final)
83%
Grant Probability
Favorable
3-4
OA Rounds
2y 6m
To Grant
99%
With Interview

Examiner Intelligence

Grants 83% — above average
83%
Career Allow Rate
260 granted / 313 resolved
+28.1% vs TC avg
Strong +20% interview lift
Without
With
+19.7%
Interview Lift
resolved cases with interview
Typical timeline
2y 6m
Avg Prosecution
23 currently pending
Career history
336
Total Applications
across all art units

Statute-Specific Performance

§101
15.9%
-24.1% vs TC avg
§103
47.8%
+7.8% vs TC avg
§102
12.1%
-27.9% vs TC avg
§112
14.3%
-25.7% vs TC avg
Black line = Tech Center average estimate • Based on career data from 313 resolved cases

Office Action

§103
DETAILED ACTION Claims 1, 3-4, 9, 11-12, 16 and 18-19 have been amended. Claims 1-20 remain pending in the application. Claims 1, 9 and 16 are independent. The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office 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 . Response to Amendment and Arguments Applicant's arguments regarding rejection under 35 U.S.C. § 103 have been fully considered but in moot in view of new ground of rejection. Applicant amended independent claims 1, 9 and 16 to further specify: generate a feedback signal, based on the generated score, and use the feedback signal to update the model. ZHENG CN 111079540 A is introduced in view of new ground of rejection. The teachings of Phan and MULLIGAN as disclosed in the previous office action are hereby incorporated by references to the extent applicable to the amended claims. Another iteration of claim analysis has been made. Referring to the corresponding sections of the claim analysis below for details. 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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Phan “Regression Optimization for System-level Production Control” 2021 in view of MULLIGAN US 20210057098 A11 and ZHENG CN 111079540 A. Regarding claim 1, Phan teaches method steps: obtain a plurality of regression functions that predict an output of a plurality of processes of a physical system based on inputs received at each process (page 5024 Fig. 2 left column second paragraph from the bottom to right column paragraph 2, a circular node representing a process with regression function for modeling output of the process is built based on historical data); automatically generate one or more constraints for a model for the physical system based at least in part on the plurality of regression functions and a representation of the physical system, wherein the representation specifies relationships between at least a portion of the plurality of processes (page 5024 Fig. 2 right column paragraphs 2-3, generating linear constraint based on the process output defined by the regression functions i.e. “automatically generate one or more constraints” and defining objective function based on the regression function for a optimization model for the plant, the plant is modeled using process nodes with output prediction model based on regression function and operational constraint nodes i.e. “a model for the physical system based at least in part on the plurality of regression functions and a representation of the physical system”); identify a set of parameter values for controlling the physical system based on the model (Fig. 1 page 5023 – 5024 paragraphs 1-4 in section II, optimal set points for the plant process are derived based on the optimization model) by minimize the objective function (page 5024 Fig. 2 right column paragraphs 2-3, find optimal set points to minimize objective function); and cause the physical system to be configured in accordance with the set of parameter values (Fig. 1 page 5023 – 5024 paragraphs 1-4 in section II, optimal set points for the plant process and devising set point trajectories over a time horizon). Phan does not explicitly teach: the method steps are implemented by a system comprising a memory configured to store program instructions; a processor operatively coupled to the memory; generate a score, for the set of parameter values, based on a predicted improvement of a confidence bound of the set of parameter values for controlling the physical system relative to historical performance of the physical system; in response to the generated score satisfying a threshold, complete the optimization; automatically generate one or more objective functions for the model for the physical system; and generate a feedback signal, based on the generated score, and use the feedback signal to update the model. MULLIGAN explicitly teaches in an analogous art that: the method steps are implemented by a system comprising a memory configured to store program instructions; a processor operatively coupled to the memory (Fig. 1 [0061] computer system/server 12); generate a score, for the set of parameter values, based on a predicted improvement of a confidence bound of the set of parameter values for controlling the physical system relative to historical performance of the physical system; and in response to the generated score satisfying a threshold, complete the optimization ([0081] [0094] [0096] computes a core based on a confidence interval using a threshold on the score, the threshold value is dynamically learned based on historical data, i.e. “generate a score, for the set of parameter values, based on a predicted improvement of a confidence bound of the set of parameter values for controlling the physical system relative to historical performance of the physical system”, the optimum actions are added upon the score exceeding the threshold, i.e. complete the optimization); ZHENG explicitly teaches in an analogous art that: automatically generate one or more objective functions for the model for the physical system (page 3 paragraphs 1-4, and paragraphs 11-14, constructing loss function i.e. “automatically generate one or more objective functions”); and generate a feedback signal, based on the generated score, and use the feedback signal to update the model (Fig. 1, paragraphs 10-17, calculating average detection precision mAP index of the model, if the mAP index does not satisfy the requirement, updating the neural network structure iteratively). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Phan to incorporate the teachings of MULLIGAN and ZHENG, because they all directed to process optimization using objective function, to make the system wherein the method steps are implemented by a system comprising a memory configured to store program instructions; a processor operatively coupled to the memory; generate a score, for the set of parameter values, based on a predicted improvement of a confidence bound of the set of parameter values for controlling the physical system relative to historical performance of the physical system; in response to the generated score satisfying a threshold, complete the optimization; the one or more objective functions are generated; and generate a feedback signal, based on the generated score, and use the feedback signal to update the model. One of ordinary skill in the art would have been motivated to do this modification so as to recommend one or more optimal actions, as MULLIGAN teaches in [0089], and to optimize the model until the index meeting the requirement, as ZHENG teaches in page 4 paragraph 10. Regarding claim 2, Phan further teaches the regression functions are automatically generated and obtained from a machine learning framework (page 5024 left column paragraphs 1-2 and right column paragraphs 1-2, the regression function modeling process output are built by trained machine leaning model). Regarding claim 3, ZHENG further teaches use the feedback signal to update the machine learning framework (paragraphs 10-17, calculating average detection precision mAP index of the model, if the mAP index does not satisfy the requirement, updating the neural network structure). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Phan to incorporate the teachings of ZHENG, because they all directed to process optimization using objective function, to make the system wherein use the feedback signal to update the machine learning framework. One of ordinary skill in the art would have been motivated to do this modification so as to optimize the model until the index meeting the requirement, as ZHENG teaches in page 4 paragraph 10. Regarding claim 4, ZHENG further teaches in response to the generated score not satisfying the threshold, identify a new set of parameter values based on one or more of the updated model and the updated machine learning framework (Fig. 1, paragraphs 10-17, calculating average detection precision mAP index of the model, if the mAP index does not satisfy the requirement, updating the neural network structure iteratively, i.e. repeating the optimization process). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Phan to incorporate the teachings of ZHENG, because they all directed to process optimization using objective function, to make the system wherein in response to the generated score not satisfying the threshold, identify a new set of parameter values based on one or more of the updated model and the updated machine learning framework. One of ordinary skill in the art would have been motivated to do this modification so as to optimize the model until the index meeting the requirement, as ZHENG teaches in page 4 paragraph 10. Regarding claim 5, Phan further teaches the representation comprises a directed graph, wherein the plurality of processes of the physical system is represented as nodes in the directed graph, and wherein the relationships between at least a portion of the plurality of processes are represented as edges in the directed graph (Fig. 2 left column second paragraph from the bottom to right column paragraph 2, a circular node representing a process, the segments with arrow i.e. “edge” represents the relationships between at least a portion of the plurality of processes). Regarding claim 6, Phan further teaches the directed graph comprises one or more cycles (Fig. 2 the circular nodes). Regarding claim 7, Phan further teaches the physical system corresponds to a manufacturing plant that produces one or more products (page 5023 first paragraph of Section II, the oil sands processing plant). Regarding claim 8, Phan further teaches the set of parameter values specify a configuration for each of the plurality of processes (Fig. 1 page 5024 left column paragraph 1, mine tonnage rates and upgrading feed rates). Regarding claims 9-15, they are directed to a program of carrying out the system with similar limitations as set forth in claims 1-7, respectively. Since Phan, MULLIGAN and ZHENG teach the claimed system, they teach the program for implementing the system. Regarding claims 16-20, they are directed to a method of carrying out the system with similar limitations as set forth in claims 1-5, respectively. Since Phan, MULLIGAN and ZHENG teach the claimed system, they teach the method steps for implementing the system. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. WANG CN 110784465 A teaches constructing loss functions and adjusting model parameters when performance score is lower than threshold. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Michael Tang whose telephone number is (571)272-7437. The examiner can normally be reached M-F 7:30-4 EST. 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, Kamini Shah can be reached on (571)272-2279. 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. /M.T./ Examiner, Art Unit 2115 /KAMINI S SHAH/Supervisory Patent Examiner, Art Unit 2115 1 Phan and MULLIGAN are the prior arts of record
Read full office action

Prosecution Timeline

Sep 20, 2022
Application Filed
May 03, 2025
Non-Final Rejection — §103
Aug 06, 2025
Response Filed
Sep 20, 2025
Final Rejection — §103
Dec 22, 2025
Request for Continued Examination
Jan 10, 2026
Response after Non-Final Action
Feb 04, 2026
Non-Final Rejection — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12602043
ASSET FAILURE AND REPLACEMENT MANAGEMENT OF A SET OF ASSETS
2y 5m to grant Granted Apr 14, 2026
Patent 12591215
SYSTEM AND METHOD FOR FORMING DENTAL APPLIANCES
2y 5m to grant Granted Mar 31, 2026
Patent 12585231
Digital-Twin-Enabled Artificial Intelligence System for Distributed Additive Manufacturing
2y 5m to grant Granted Mar 24, 2026
Patent 12585261
MACHINING SIMULATION DEVICE AND MACHINING SIMULATION METHOD
2y 5m to grant Granted Mar 24, 2026
Patent 12578703
SYSTEMS AND METHODS FOR ADDITIVELY MANUFACTURING SEMI-CUSTOM OBJECTS
2y 5m to grant Granted Mar 17, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

AI Strategy Recommendation

Get an AI-powered prosecution strategy using examiner precedents, rejection analysis, and claim mapping.
Powered by AI — typically takes 5-10 seconds

Prosecution Projections

3-4
Expected OA Rounds
83%
Grant Probability
99%
With Interview (+19.7%)
2y 6m
Median Time to Grant
High
PTA Risk
Based on 313 resolved cases by this examiner. Grant probability derived from career allow rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month