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 .
Final Rejection
Applicant's arguments filed 5/28/2026 have been fully considered but they are not persuasive for reasons detailed below.
The 35 U.S.C. 101 and 112 rejections are maintained or modified as follows:
These rejections have been withdrawn.
The prior art rejections are maintained or modified as follows:
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 of this title, 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, 2, 8, 9, 12, 13 and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Brouwers (US 2025/0044271) in view of El Gamal et al. (“Gamal”)(US 2021/0243226); Glas et al. (“Glas”)(US 2024/0275809) and Shumpert (US 2016/0342903).
Brouwers teaches a system and related method comprising:
(re: certain elements of claim 12) one or more processors; and a memory coupled to the one or more processors, wherein the memory is configured to provide the one or more processors with instructions (fig. 42 and para. 314-320 teaching control system including processor 993 and memory 994 configured to carry computer-executable instructions described below) which when executed cause the one or more processors to:
receive a dataset for anomaly detection (para. 60, 76, 85, 90 teaching unsupervised learning step using unlabeled datasets for training and subsequent anomaly detection);
train an unsupervised machine learning model using at least a portion of the training dataset, to generate a trained unsupervised machine learning model (Id.);
train a supervised machine learning model using an output from the unsupervised machine learning model and an anomaly detection feedback associated with the output from the unsupervised machine learning model, to generate a trained supervised machine learning model (para. 14, 77, 155, 290 teaching training a supervised model using output of trained unsupervised ML model); and
provide both the trained unsupervised machine learning model and the trained supervised machine learning model for combined use in machine learning anomaly detection inference (para. 78-79 teaching making anomaly inference, i.e., prediction of “normal” or “anomalous” using the hybrid model);
(re: claim 18) wherein the memory is further configured to provide the one or more processors with instructions which when executed cause the one or more processors to detect an anomaly by performing a machine learning anomaly detection inference process,
wherein a prediction output of the trained unsupervised machine learning model is provided as an input to the trained supervised machine learning model (para. 76-78, 88-89 teaching use of prediction output from unsupervised machine learning model as input allows for detection of anomalous behavior);
(re: claims 19) wherein the unsupervised machine learning model is trained to reconstruct an input provided to the unsupervised machine learning model, and the unsupervised machine learning model provides the reconstructed input as an input to the trained supervised machine learning model (para. 14, 89, 208, 217-218 teaching reconstructing process data and using the output of the reconstruction error as input for the supervised machine learning model).
(re: claims 1, 2, 8, 9 and 20) The claimed method steps and related computer program product are taught in the normal operation of the combined system described below.
Brouwers as set forth above teaches all that is claimed except for expressly teaching
(re: certain elements of claims 1, 12) selecting a portion of the dataset as training data responsive to determining that a data profile for the portion of the dataset corresponds to training data;
preprocessing the portion of the dataset, wherein preprocessing the portion of the dataset comprises removing data associated with anomalies; and
train a supervised machine learning model using anomaly detection feedback associated with the output from the unsupervised machine learning model;
(re: claims 2, 13) wherein the memory is further configured to provide the one or more processors with instructions which when executed cause the one or more processors to determine and provide an indication of a maturity of the supervised machine learning model;
(re: claim 20) determining a level of concept drift exhibited by the unsupervised machine learning model; and
responsive to determining that the level of concept drift falls outside of a threshold range of values, retraining the unsupervised machine learning model.
Gamal, however, teaches that it is well-known in the artificial intelligence arts
(re: certain elements of claims 1, 12)
- to use clean/normal data, i.e., anomaly-free data, to (re)train a model so that the model can learn normal behavior from a baseline data set that is uncontaminated with anomalous data to improve the accuracy of the machine learning model (fig. 1 near 114 and para. 20, 29, 34-36 teaching anomaly detection framework 114 that includes an unsupervised machine learning model 115 that is trained with a normal dataset, wherein the framework updates/retrains the machine learning models based with a continuously updated normal dataset that is kept free of anomalous data).
Glas further teaches that it is well-known in the artificial intelligence arts
(re: certain elements of claims 1, 2, 12, 13)
-to also update the machine learning model with user feedback and new training data to maintain accuracy of the model as well as the related maturity score (para. 76- 79).
Shumpert further teaches that it is well-known in the artificial intelligence arts
(re: certain elements of claims 1, 12 and 20)
to train unsupervised machine learning models with normal data and to also dynamically update the models to handle the issue of concept drift to improve model accuracy in anomaly detection and to reduce the incidences of false alarms (para. 1, 9, 13, 21, 28, 31, 34, 65 teaching use a training data set gathered under normal conditions as well as dynamically updating the models to address the naturally change of normal parameters, wherein dynamic updating is regarded as automatic updating after a certain level of drift/change has occurred; see also para. 60-64 teaching use of anomaly detection feedback to improve accuracy of machine learning models via continuous updates).
It would thus be obvious to one with ordinary skill in the art to modify the base reference with these prior art teachings—with a reasonable expectation of success—to arrive at the claimed invention. The rationale for this obviousness determination can be found in the prior art itself as cited above. Further, the prior art discussed and cited demonstrates the level of sophistication of one with ordinary skill in the art and that these modifications are predictable variations that would be within this skill level. Therefore, it would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the invention of Brouwers for the reasons set forth above.
Claims 3-7, 10 and 14-17 are rejected under 35 U.S.C. 103 as being unpatentable over Brouwers, Gamal, Glas and Shumpert (“Brouwers et al.”) as applied to the claims above, and further in view of Tora (US 2022/0277219), Delisle et al. (“Delisle”)(US 2023/02979888) and legal precedent.
Brouwers et al. as set forth above teach all that is claimed except for expressly teaching
(re: certain elements of claim 3) wherein determining the indication of the maturity of the supervised machine learning model includes performing a loss calculation;
(re: claims 4, 14) wherein the indication of the maturity of the supervised machine learning model is associated with a maturity score, and the maturity score is based on loss calculation results for three or more epochs associated with the supervised machine learning model;
(re: claims 5, 15) wherein determining the indication of the maturity includes converting a logarithmic value to a linear value;
(re: claims 6, 16) wherein the anomaly detection feedback includes one or more features identified by a user as contributing to a predicted anomaly;
(re: claims 7, 17) wherein the supervised machine learning model is trained at a first training rate that is greater than a second training rate at which the unsupervised machine learning model is trained;
(re: claim 10) wherein the supervised machine learning model is further trained to predict a severity score of a predicted anomaly.
Delisle, however, teaches that it is well known in the machine learning arts to calculate maturity scores using various methods—including a logarithmic loss function—and that the number of epochs in calculations can be varied (para. 68-69).
Tora further teaches that it is well known in the machine learning arts to adjust the learning rate and number of epochs in loss calculations and that the user can configure the various features and variables of the ML algorithm to test the received data (para. 30-32 teaching that learning rate and the number of epochs are well-known hyperparameters when configuring models and that a combination of supervised as well as non-supervised human-tagged or classified data can be used; para. 33 teaching user interface that allows user to select various features and configuration parameters—including what type of ML model).
Indeed, the claimed features relating to type of ML model, learning rates, number of epochs, severity and maturity scores can be regarded as common design parameters/operating variables controlled by the design incentives and/or economic considerations involved in this type of subject matter. This is especially applicable in the machine learning arts as demonstrated above. Moreover, legal precedent teaches that variations in these type of common design parameters/operating variables are obvious and are the mere optimization of result-effective variables that would be known to one with ordinary skill in the art. See MPEP 2144.05 I.II (teaching ample motivation to optimize or modify result-effective variables based on “design need(s)” or “market demand”).
It would thus be obvious to one with ordinary skill in the art to modify the combination of references with these prior art teachings—with a reasonable expectation of success—to arrive at the claimed invention as these modifications are already well-known and commonly implemented in the machine learning arts. The rationale for this obviousness determination can be found in the prior art itself as cited above and in legal precedent as described above.
Further, the prior art discussed and cited demonstrates the level of sophistication of one with ordinary skill in the art and that these modifications are predictable variations that would be within this skill level. Therefore, it would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the invention of Brouwers et al. for the reasons set forth above.
Response to Arguments
Applicant’s arguments that the prior art fails to teach the claim features are unpersuasive in view of the reformulated prior art rejection set forth above. In particular, the rejection explains that the preparation of a training data set with normal data, i.e., uncontaminated with anomalous data, is well-known in the machine learning arts for improving the accuracy of a machine learning model. Consequently, as a reasonable interpretation of the prior art renders Applicant’s claimed invention obvious, the claims stand rejected.
Allowable Subject Matter
Claim 11 is 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.
Examiner has maintained the prior art rejections, statutory rejections and drawing objections as previously stated and as modified above. Applicant's amendment necessitated any new grounds 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). The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
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 extension fee 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 date of this final action.
Conclusion
Any references not explicitly discussed but made of record during the prosecution of the instant application are considered helpful in understanding and establishing the state of the prior art and are thus relevant to the prosecution of the instant application.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOSEPH C RODRIGUEZ whose telephone number is 571-272-3692 (M-F, 9 am – 6 pm, PST). The Supervisory Examiner is MICHAEL MCCULLOUGH, 571-272-7805.
Alternatively, to contact the examiner, send an E-mail communication to Joseph.Rodriguez@uspto.gov. Such E-mail communication should be in accordance with provisions of the MPEP (see e.g., 502.03 & 713.04; see also Patent Internet Usage Policy Article 5). E-mail communication must begin with a statement authorizing the E-mail communication and acknowledging that such communication is not secure and may be made of record. Please note that any communications with regards to the merits of an application will be made of record. A suggested format for such authorization is as follows: "Recognizing that Internet communications are not secure, I hereby authorize the USPTO to communicate with me concerning any subject matter of this application by electronic mail. I understand that a copy of these communications will be made of record in the application file”.
Information regarding the status of an application may also be obtained from the Patent Center: https://patentcenter.uspto.gov/
/JOSEPH C RODRIGUEZ/Primary Examiner, Art Unit 3655
Jcr
------
August 11, 2026
Prosecution Insights