DETAILED 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. Status of the Claims Claims 1 - 2, 4 - 8, 10-1 2 , 14 - 17, and 19-20 are rejected under 35 U.S.C. 102(a)(1). Claims 3, 9, 13, and 18 are rejected under 35 U.S.C. 103. Specification The title of the invention is not descriptive. A new title is required that is clearly indicative of the invention to which the claims are directed. Claim Rejections - 35 USC § 102 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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action: A person shall be entitled to a patent unless – (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale , or otherwise available to the public before the effective filing date of the claimed invention. Claims 1 - 2, 4 - 8, 10- 1 2 , 14 - 17, and 19- 20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by PATHAK (US 2022/0123552 A1). Regarding Claim 1 , PATHAK teaches a computer-implemented method comprising: receiving operational data representing operations of an asset; (¶ 30: An asset in the reference is a power network of a factory, plant, building, or other facility , which is similar to the example of “an asset” in applicant’s disclosure in paragraph 53-54 . ¶ 37: Operational data is provided to an anomaly detection system, which includes information or data regarding live power consumption, processes, production, and environment associated with a facility , i.e. an asset .) processing the operational data to generate a fault anomaly score for the operational data; (¶ 38, 40: “ The anomaly detection module 224 identifies potential anomalies in the power data and provides this information to an anomaly scoring module 226. ” See Fig. 3A steps 314-324: An anomaly score is calculated by an anomaly detection module using the stream of operation data. ) determining, based at least in part on the fault anomaly score, whether the operational data is indicative of the asset being associated with at least one fault; in accordance with a determination that the operational data is indicative of the asset being associated with the at least one fault: (¶ 38, 47: Fig. 3A step 324 to Fig. 3C step 336: After calculating the anomaly score, the score is evaluated to determine if it is redundant, and if it is not redundant (i.e. it is a fault), the type and severity of the fault is to be classified. ) generating, based at least in part on applying the operational data to a fault classification model, fault data; (¶ 46-47, Fig. 3C step 338: A trained classification model 358 is used to determine the type and severity of the anomaly, which is fault data. ) generating, based at least in part on applying the fault data to a fault impact model, fault impact data , wherein the fault impact model comprises a reinforcement learning model ; (¶ 43-44, 47, Fig. 3C steps 348-352, 362 , 364 to Fig. 3B step 328: Based at least in part on fault data , including type and severity of the fault, and a corresponding corrective action, expected changes to the power data (i.e. fault impact data) are determined. If the data changed as expected, positive feedback is provided. If it did not, negative feedback is provided. The impact on the power data is fed back to the classification model, which determines both the fault type and severity and the corrective actions to be made in a feedback loop. This process reads on “reinforcement learning” , as understood by a person of ordinary skill in the art, since the positive (i.e. reward) and negative (i.e. penalty) feedback values are responsive to an action being taken and the feedback values are “ accumulated ” and used to further train the model until the accumulated feedback values are improved . ) and initiating performance of one or more fault optimization actions based at least in part on the fault impact data. (¶ 47, Fig. 3C: “ The alarm notification may include suggested action items, at block 344, based on historical fault information and repair actions that were undertaken , which are stored in a fault database 360 … the anomaly detection continues to look for expected changes in the power data responsive to the corrective action that was taken , at block 348. Based on the [ co rr ective ] action taken and subsequent expected changes, at block 350, positive feedback or negative feedback are provided such that implicit and explicit feedback may be respectively accumulated, at block 352. These implicit and/or explicit feedbacks may then be provided to trigger the training pipeline, at block 328 ” . Also see ¶ 30, 32: Corrective actions are taken that mitigate faults or increase performance of elements in a facility and are therefore “fault optimization actions”. The actions are recommended and then performed based on the output of the classification model, which is trained using a reinforcement learning feedback loop comprising the impact of the fault and the corrective action on the power data. An action in one iteration of the loop is therefore at least in part based on fault impact data. ) Claim 11 is directed to an apparatus and Claim 20 is directed to a computer program product but they otherwise recite the same limitations as claim 1. Claim 11 and claim 20 are therefore rejected Regarding Claim 2 , PATHAK further teaches wherein the one or more fault optimization actions include at least one short-term fault optimization action. (¶ 30-31, 39: Anomalies are diagnosed in real-time in order to take corrective actions. The corrective actions are therefore short-term fault optimization actions.) Claim 12 recites the same limitations as claim 2 and is therefore rejected for the same reasoning discussed above. Regarding Claim 4 , PATHAK further teaches further comprising: performing a first training of the fault impact model, wherein the first training of the fault impact model comprises: identifying a historical dataset, wherein the historical dataset comprises labeled data; and training the fault impact model using a machine learning technique based at least in part on the historical dataset. (¶ 37, 42, 46: The classification model classifies the faults, their impact, and their respective actions used to remedy or repair the faults. The first training is the pre-training of the classification model. The pre-training involves accessing the stored historical faults and their mappings to their remedies. These mappings are labels used in training the machine learning model. ) Claim 1 4 recites the same limitations as claim 4 and is therefore rejected for the same reasoning discussed above. Regarding Claim 5 , PATHAK in view of NIXON further teaches wherein the machine learning technique is a supervised machine learning technique. (¶ 42-43, 46: The user provides feedback that is used in the adaptive learning to train the model. The machine learning model is also trained using known anomaly types and severity levels from historical records and their mapping to remedy actions. This is labeled data. Therefore, the machine learning technique is a supervised machine learning technique. ) Claim 1 5 recites the same limitations as claim 5 and is therefore rejected for the same reasoning discussed above. Regarding Claim 6 , PATHAK further teaches further comprising: performing a second training of the fault impact model, wherein the second training of the fault impact model comprises: training the fault impact model based at least in part on the fault data. ( ¶ 43, 47, Fig. 3C: step 362: The classification model, which also determines the impact of the fault and the recommended action to be performed, is retrained based at least in part by the fault data that is updated in the fault database, including the severity and the type of the fault, which is fed back to the model both explicitly via user feedback (labeled) and implicitly by determining if an expected change in the operational data is detected (unlabeled) . ) Regarding Claim 7 , PATHAK in view of NIXON further teaches wherein the fault data is unlabeled data. (¶ 43, 47: The implicit feedback about the fault data and the corrective action that was taken is unlabeled data, i.e. data that is not provided by a user or otherwise indicated as being the correct or incorrect outcome .) Claim 1 6 recites the same limitations as claim 7 and is therefore rejected for the same reasoning discussed above. Regarding Claim 8 , PATHAK further teaches wherein the fault data indicates a fault type associated with the at least one fault. (¶ 46-47, 51, 67, Fig. 3C step 338: The classified fault data includes a fault type and severity.) Claim 1 7 recites the same limitations as claim 8 and is therefore rejected for the same reasoning discussed above. Regarding Claim 10 , PATHAK further teaches wherein the at least one fault is at least one of a transient fault, an intermittent fault, or a permanent fault. (¶ 49, Figs. 5A-5C: An example fault includes at least one of a transient or intermittent fault, as illustrated by the transient or intermittent increase or decrease in operational data in the graphs of Figs. 5A-5C. ) Claim 1 9 recites the same limitations as claim 10 and is therefore rejected for the same reasoning discussed above. 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 . Claims 3 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over PATHAK ( US 2022 / 0123552 A1 ) in view of ZHOU ( US 2024 / 0144052 A1 ). Regarding Claim 3 , PATHAK teaches all the limitations of claim 1, on which claim 3 depends. PATHAK does not teach wherein the one or more fault optimization actions include at least one long-term fault optimization action. However, ZHOU, which is similarly directed to fault detection and remediation for maintaining physical assets, teaches wherein the one or more fault optimization actions include at least one long-term fault optimization action. (¶ 25-26, 57-59, 63: An optimization model for maintaining a physical asset outputs a plan or schedule including maintenance actions, such as repairing or replacing the asset, to be performed. A scheduled maintenance plan is a long-term fault optimization action. ) Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to modify the recommendation and performance of remedial actions responsive to detecting a fault in an asset taught by PATHAK by including long-term remedial actions, such as a scheduled maintenance plan , as taught by ZHOU. Since the references are similarly directed to optimizing the performance of physical assets and of using models to determine a maintenance action, the combination would have yielded predictable results. As taught by ZHOU ( ¶ 25 ) , automating the determination of long-term maintenance actions would enhance productivity and improve the scalability and adaptability of the model to different maintenance plan time horizons. Claim 1 3 recites the same limitations as claim 3 and is therefore rejected for the same reasoning discussed above. Claims 9 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over PATHAK ( US 2022 / 0123552 A1 ) in view of NIXON ( US 2024 / 0119342 A1 ) . Regarding Claim 9 , PATHAK teaches all the limitations of claim 1, on which claim 9 depends. PATHAK does not explicitly teach wherein the fault anomaly score is generated at least in part by performing a principal component analysis technique. However, NIXON, which is similarly directed to fault detection and optimization using reinforcement learning (¶ 15), teaches wherein the fault anomaly score is generated at least in part by performing a principal component analysis technique. ( ¶ 64, 66-67, Fig. 3: A fault or anomaly is detected by using a dynamic threshold calculated using principal component analysis and updated using reinforcement learning. ) Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to modify the detection of a fault using an anomaly score taught by PATHAK by using , at least in part, a principal component analysis technique in the anomaly detection as taught by NIXON. Since the references are similarly directed to fault detection and remedying in industrial environments using reinforcement learning , the combination would have yielded predictable results. As taught by NIXON (¶ 9 ), use of PCA is a conventional solution for fault detection by a threshold, or score ; however, it can be improved using reinforcement learning to reduce the number of false positives (¶ 15). Inclusion of a PCA technique in determining the fault detection score taught by NIXON while incorporating the machine learning techniques taught by both of the references would have therefore been advantageous to a person or ordinary skill in the art to improve the accuracy of the fault detection and classification. Claim 1 8 recites the same limitations as claim 9 and is therefore rejected for the same reasoning discussed above. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Picos ( US 2022 / 0198054 A1 ) teaches machine learning techniques for detecting an anomaly, determining a risk score, and inferring a remediation to be performed to mitigate the anomaly. (¶ 77) Bello ( US 2020 / 0233397 A1 ) teaches continuously generating industrial machine health metrics using machine learning models to determine the probability of a detection of a range of fault types at varying levels of severity . (¶ 43) DePalo ( US 12 , 341 , 791 B1 ) teaches determining an initial score for anomaly detection and then using a machine learning model to further classify the anomaly using the operational timeseries data. (Col. 4:50-67) Any inquiry concerning this communication or earlier communications from the examiner should be directed to FILLIN "Examiner name" \* MERGEFORMAT RAMI RAFAT OKASHA whose telephone number is FILLIN "Phone number" \* MERGEFORMAT (571)272-0675 . The examiner can normally be reached FILLIN "Work Schedule?" \* MERGEFORMAT M-F 10-6 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, FILLIN "SPE Name?" \* MERGEFORMAT SCOTT BADERMAN can be reached at FILLIN "SPE Phone?" \* MERGEFORMAT (571) 272-3644 . 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. /RAMI R OKASHA/ Primary Examiner, Art Unit 2118