Notice of Pre-AIA or AIA Status
This Final communication is in response to Application No. 18/013,894 filed on 12/29/2022 which claims priority to EP20183021.3 filed on 06/30/2020. 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
The amendment filed 1/22/2026 which provides amendments to claims 1, 9, 11, 13 and cancels claim 18. Claims 1-11, 13-17 and 19 are pending. Amendments to the claims has overcome the 112(d) rejection.
Response to Arguments
Applicant’s arguments with respect to 35 U.S.C § 112(f) filed 1/22/2026 (page 10 of applicant’s arguments) have been fully considered but they are not persuasive.
Applicant argues that the terms “first interface”, “computation unit” and “second interface” do not invoke 112(f) and recites that [0029] and [0045] of the specification show that these terms are defined to include structure. The examiner respectfully disagrees because the specification provides examples of these claim limitations (For example [0029] “computation unit 124 included in the processing system 100 may correspond to the claimed system (e.g., IT system), which may include one or more suitably configured processors and memory.”) and does not provide an explicit definition for these terms. Thus, they invoke 112(f).
Applicant’s arguments with respect to 35 U.S.C § 112(b) filed 1/22/2026 (top of page 11 of applicant’s arguments) have been fully considered but they are not persuasive.
Applicant argues that the term “given respective anomaly score” does not lack antecedent basis as there is a distinction between “determined respective anomaly score” and “given anomaly score”. The examiner respectfully disagrees. The “given anomaly score” is not mention previously in the claim and thus it is unclear how these “given anomaly score” are provided. Are they stored in memory, or provided from a user? Claim 4 recites them being determined but that makes it seem like the “given anomaly score” is determined after they have already been used. Thus, the 112(b) rejection is maintained.
Applicant’s arguments with respect to 35 U.S.C § 101 filed 1/22/2026 (pages 11-13 of applicant’s arguments) have been fully considered but they are not persuasive.
Applicant argues that the claimed invention is an improvement in anomaly detection saying that it provides the framework that identifies why the data is different rather than identifying that the data is different (page 12 of applicant’s argument). The examiner respectfully disagrees because the claimed invention is targeted towards determining anomaly scores for N classes and comparing these determined anomaly scores to given anomaly scores. Finally an alarm is provided when the difference is above a threshold. It is unclear how this shows why the data is different.
Applicant also argues that characterizing the “determining” step as a “Mental Process” is improper (page 13 of applicant’s argument) . The examiner respectfully disagrees. One of the “determining” steps is directed to determining anomaly scores for a data batch. Given the broadest reasonable interpretation of this limitation this could be done in the mind of a human. Having machine learning models perform the task does not make it any less of a mental process. The other “determining” step is directed to comparing anomaly scores. Also, given the broadest reasonable interpretation of this limitation this could be done in the mind of a human. Thus, the 101 rejection is maintained.
Applicant’s arguments with respect to 35 U.S.C § 103 filed 1/22/2026 (pages 13-19 of applicant’s arguments) have been fully considered but they are not persuasive.
Applicant argues that the prior art does not teach “the at least N separable classes relate to typical scenarios of the respective device and correspond to a state of proper operation and to N-1 typical failure modes of the respective device”, “The N anomaly detection model include N trained anomaly detection models, the N trained anomaly detection models being trained using a reference data set that is provided beforehand by identifying the typical scenarios and related to typical variables or input data” and “ the typical scenarios include a scenario when the respective device is working properly, and when the respective device is damaged” (pages 14-18 of applicant argument). The examiner respectfully disagrees as Ardel teaches all of these elements. Ardel teaches operating states which comprise a “normal” state and an “abnormal” state. This is teaching the claimed typical scenarios. Ardel also trains their anomaly detection models based on historic data. More detail is provided in the 103 rejection below. Thus, the 103 rejection is maintained.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) in claim 11:
a first interface configured to
a computation unit configured to
a second interface, configured to
The first interface, computation unit, and second interface are generic placeholders without definite structure, and is described in purely functional terms using “configured to” language rather than reciting how the functions are performed. As a result, it is interpreted as means-plus-function limitation under 35 U.S.C 112(f).
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 1-11 & 13-19 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1, 11 and 13 recites the limitation "given respective anomaly scores". There is insufficient antecedent basis for this limitation in the claim.
Claims 2-10, which are dependent on claim 1, and claims 14-19, which are dependent on claim 13, incorporate the rejections of claim 1 and 13 as well.
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 1-11 & 13-19 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
101 Subject Matter Eligibility Analysis
Step 1: Claims 1-11 & 13-19 are within the four statutory (a process, machine, manufacture or composition of matter.) Claims 1-10 describe a process and 11, 13-19 describes a machine.
With respect to claim 1:
Step 2A Prong 1: The claim recites an abstract idea enumerated in the 2019 PEG
determining respective anomaly scores for the respective incoming data batch relating to the at least N separable classes(This is an abstract idea of a "Mental Process." The "determining" step under its broadest reasonable interpretation, covers concepts that can be practically performed in the human mind. The determination could be made manually by an individual.)
determining, for the respective incoming data batch, a difference between the determined respective anomaly scores for the at least N separable classes and given respective anomaly scores of the N anomaly detection models and (This is an abstract idea of a "Mental Process." The "determining" step under its broadest reasonable interpretation, covers concepts that can be practically performed in the human mind. The determination could be made manually by an individual.)
Step 2A Prong 2: The judicial exception is not integrated into a practical application
Additional elements:
receiving input data relating to at least one device, wherein the input data comprises incoming data batches relating to at least N separable classes, with n
∈
1, ..., N; (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
using N anomaly detection models; (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
generating output data, the generating of the output data comprising applying the N anomaly detection models to the input data, the output data being suitable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective device (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
when the respective determined difference between is greater than a difference threshold, providing an alarm relating to the determined difference to a user, the respective device: an IT system connected to the respective device, or any combination thereof. (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
wherein the at least N separable classes relate to typical scenarios of the respective device and correspond to a state of proper operation and to N-1 typical failure modes of the respective device, (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
wherein the N anomaly detection models include N trained anomaly detection models, the N trained anomaly detection models being trained using a reference data set that is provided beforehand by identifying the typical scenarios and related to typical variables or input data, and (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
wherein the typical scenarios include a scenario when the respective device is working properly, and when the respective device is damaged. (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional elements “receiving…”, “when the respective…”, “wherein the at least N separable classes…” and “wherein the typical scenarios…” add insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
The additional elements “generating output data…”, “using N anomaly models” and “wherein the N anomaly detection models…” are recited in a generic level and they represent generic computer components to apply the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept (MPEP 2106.05(f)).
When considered in combination, these additional elements represent insignificant extra-solution activity and mere instructions to apply an expectation, which do not provide an inventive concept.
Therefore, claim 1 is ineligible.
With respect to claim 2:
Step 2A Prong 1: claim 2, which incorporates the rejection of claim 1, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
the input data undergoes a distribution drift involving an increase of the determined difference. (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element adds insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
Therefore, claim 2 is ineligible.
With respect to claim 3:
Step 2A Prong 1: claim 3, which incorporates the rejection of claim 1, recites an additional abstract idea:
determining a distribution drift of the input data a second difference between the anomaly scores of an earlier incoming data batch and the anomaly scores of a later incoming data batch is greater than a second threshold; and (This is an abstract idea of a "Mental Process." The "determining" step under its broadest reasonable interpretation, covers concepts that can be practically performed in the human mind. The determination could be made manually by an individual.)
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
providing a report relating to the determined distribution drift to a user, the respective device: an IT system connected to the respective device when the determined difference is greater than a threshold. (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element adds insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
Therefore, claim 3 is ineligible.
With respect to claim 4:
Step 2A Prong 1: claim 4, which incorporates the rejection of claim 1, recites an additional abstract idea:
determining the given anomaly scores of the at least N separable classes for the N anomaly detection models. (This is an abstract idea of a "Mental Process." The "determining" step under its broadest reasonable interpretation, covers concepts that can be practically performed in the human mind. The determination could be made manually by an individual.)
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
assigning training data batches to the at least N separable classes of the anomaly detection models; and (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element adds insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
Therefore, claim 4 is ineligible.
With respect to claim 5:
Step 2A Prong 1: claim 5, which incorporates the rejection of claim 1, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
N=1 (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element adds insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
Therefore, claim 5 is ineligible.
With respect to claim 6:
Step 2A Prong 1: claim 6, which incorporates the rejection of claim 1, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
embedding the N anomaly detection models in a software application for analyzing, monitoring, operating, controlling, or any combination thereof of the at least one device; and (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
deploying the software application on the at least one device or an IT system connected to the at least one device: such that the software application is usable for analyzing, monitoring, operating: controlling, or any combination thereof of the at least one device (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional elements are recited in a generic level and they represent generic computer components to apply the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept (MPEP 2106.05(f)).
Therefore, claim 6 is ineligible.
With respect to claim 7:
Step 2A Prong 1: claim 7, which incorporates the rejection of claim 6, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
when the determined difference is greater than the difference threshold: amending the respective anomaly detection models, such that a determined difference using the respective amended anomaly detection models is smaller than the difference threshold; (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
replacing the respective anomaly detection models with the respective amended anomaly detection models in the software application; (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
deploying the amended software application on the at least one device or the IT system. (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional elements are recited in a generic level and they represent generic computer components to apply the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept (MPEP 2106.05(f)).
Therefore, claim 7 is ineligible.
With respect to claim 8:
Step 2A Prong 1: claim 8, which incorporates the rejection of claim 6, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
when the amendment of the anomaly detection models takes more time than a duration threshold: replacing the deployed software application with a backup software application; and (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
analyzing, monitoring, operating, controlling, or any combination thereof of the at least one device using the backup software application (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element is recited in a generic level and they represent generic computer components to apply the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept (MPEP 2106.05(f)).
Therefore, claim 8 is ineligible.
With respect to claim 9:
Step 2A Prong 1: claim 9, which incorporates the rejection of claim 1, recites an additional abstract idea:
determining a respective difference of the respective anomaly detection models; and (This is an abstract idea of a "Mental Process." The "determining" step under its broadest reasonable interpretation, covers concepts that can be practically performed in the human mind. The determination could be made manually by an individual.)
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
embedding respective N detection models in a respective software application for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected devices; (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
deploying the respective software application on the respective interconnected devices or an IT system connected to the plurality of interconnected devices, such that the respective software application is usable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected devices; (This amounts to no more than mere instructions to “apply” the exception using a generic computer component.)
providing an alarm relating to the determined difference and the respective interconnected devices for which the corresponding respective software application used for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected devices to a user, the respective device, an automation system, or any combination thereof. (this limitation amounts to adding insignificant extra-solution activity to the judicial exception).
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional elements “embedding…” and “deploying…” are recited in a generic level and they represent generic computer components to apply the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept (MPEP 2106.05(f)).
The additional element “providing…” adds insignificant extra-solution activity to the judicial exception and cannot provide an inventive concept. Storing and retrieving information in memory is directed to a well understood routine conventional activity of data transmission (MPEP 2106.05(d)(II)(iv)).
When considered in combination, these additional elements represent insignificant extra-solution activity and mere instructions to apply an expectation, which do not provide an inventive concept.
Therefore, claim 9 is ineligible.
With respect to claim 10:
Step 2A Prong 1: claim 10, which incorporates the rejection of claim 1, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
the respective device is a production machine, an automation device, a sensor, a production monitoring device, a vehicle, or any combination thereof. (this limitation merely limits the judicial exception to a particular field of use.)
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element merely limits the judicial exception to a particular field of use and cannot provide an inventive concept (MPEP 2106.05(h)).
Therefore, claim 10 is ineligible.
With respect to claim 11:
The claim recites similar limitations as corresponding to claim 1. Therefore, the same subject matter analysis that was utilized for claim 1, as described above, is equally applicable to claim 11. Therefore, claim 11 is ineligible.
With respect to claim 13:
The claim recites similar limitations as corresponding to claim 1. Therefore, the same subject matter analysis that was utilized for claim 1, as described above, is equally applicable to claim 13. Therefore, claim 13 is ineligible.
With respect to claim 14:
The claim recites similar limitations as corresponding to claim 2. Therefore, the same subject matter analysis that was utilized for claim 2, as described above, is equally applicable to claim 14. Therefore, claim 14 is ineligible.
With respect to claim 15:
The claim recites similar limitations as corresponding to claim 3. Therefore, the same subject matter analysis that was utilized for claim 3, as described above, is equally applicable to claim 15. Therefore, claim 15 is ineligible.
With respect to claim 16:
The claim recites similar limitations as corresponding to claim 4. Therefore, the same subject matter analysis that was utilized for claim 4, as described above, is equally applicable to claim 16. Therefore, claim 16 is ineligible.
With respect to claim 17:
The claim recites similar limitations as corresponding to claim 5. Therefore, the same subject matter analysis that was utilized for claim 5, as described above, is equally applicable to claim 17. Therefore, claim 17 is ineligible.
With respect to claim 19:
Step 2A Prong 1: claim 19, which incorporates the rejection of claim 1, does not recite an abstract idea.
Step 2A Prong 2: The judicial exception is not integrated into a practical application.
The system is an IT system. (this limitation merely limits the judicial exception to a particular field of use.)
Step 2B: the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception
The additional element merely limits the judicial exception to a particular field of use and cannot provide an inventive concept (MPEP 2106.05(h)).
Therefore, claim 19 is ineligible.
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.
Claims 1-7, 9-11 and 13-19 are rejected under 35 U.S.C. 103 as being unpatentable over Ardel (US 20210326759) in view of Schierz (WO 2020124037)
Regarding claim 1, Ardel teaches:
A computer-implemented method comprising: ([0005] “The present disclosure describes systems and methods that enable use of trained models to automatically detect anomalous behavior of devices belonging to the rotational equipment category.”)
receiving input data relating to at least one device, wherein the input data comprise comprises incoming data batches relating to at least N separable classes, with nϵ1, . . . , N; ([0010] “In some aspects, a self-training sensor apparatus includes a sensor to generate sensor data associated with operation of a device and an input interface to receive input to select between a training mode of operation and a detection mode of operation. The apparatus also includes a memory including model generation instructions and one or more processors coupled to the memory. The one or more processors are configured to, while operating in the training mode of operation, receive first sensor data from the sensor and execute the model generation instructions based on the first sensor data to generate a trained anomaly detection model and an alert detection model. The one or more processors are also configured to while operating in the detection mode of operation, receive second sensor data from the sensor, use the trained anomaly detection model to generate an anomaly score, and process the anomaly score using the alert detection model to determine whether to generate an alert. The apparatus further includes an output interface to output an alert indication responsive to the one or more processors generating the alert.”)
determining respective anomaly scores for the respective incoming data batch relating to the at least N separable classes using N anomaly detection models; ([0034] “The one or more processors 120 are configured to process the frequency domain data 124 using the trained anomaly detection model 140, and the trained anomaly detection model 140 is configured to generate an anomaly score 142 for each sensing period.”)
determining, for the respective incoming data batch, a difference between the determined respective anomaly scores for the at least N separable classes and given respective anomaly scores of the N anomaly detection models; and ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
when the respective determined difference between is greater than a difference threshold, providing an alarm relating to the determined difference to a user, the respective device, an IT system connected to the respective device, or any combination thereof. ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
wherein the at least N separable classes relate to typical scenarios of the respective device and correspond to a state of proper operation and to N-1 typical failure modes of the respective device, ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
wherein the N anomaly detection models include N trained anomaly detection models, the N trained anomaly detection models being trained using a reference data set that is provided beforehand by identifying the typical scenarios and related to typical variables or input data, and ([0005] “In some implementations, the models can be automatically generated and trained for each particular asset based on historic non-anomalous data for that particular asset.”)
wherein the typical scenarios include a scenario when the respective device is working properly, and when the respective device is damaged. ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
Ardel does not teach generating output data, the generating of the output data comprising applying the N anomaly detection models to the input data, the output data being suitable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective device. But Schierz does ([00104] “In some embodiments, an anomaly detection model simply provides a binary prediction of anomalousness of a data sample. For example, in some embodiments, an anomaly detection model can output a prediction of a data sample as anomalous or non-anomalous. In alternative embodiments, an anomaly detection model can generate an anomaly score for a data sample. An anomaly score for a data sample may indicate an extent to which the data sample is anomalous. In certain embodiments the extent to which a data sample is anomalous can be a probability that the data sample is anomalous and/or a degree to which the data sample is anomalous. For example, an anomaly detection model may generate an anomaly score of 0.9 for a data sample, thereby predicting that there is a 90% chance that the data sample is anomalous.”).
Ardel and Schierz are considered analogous art to the claimed invention because they are in the
same field of endeavor being anomaly detection. It would have been obvious to one of
ordinary skill in the art before the effective filing date of the claimed invention to combine the anomaly detection system/method of Ardel with the anomaly detection method of Schierz. One would want to do this to achieve more accurate anomaly detection.
Regarding claim 2, Ardel in view of Schierz teaches claim 1 as outlined above. Schierz further teaches:
the input data undergoes a distribution drift involving an increase of the determined difference. ([00192] “To avoid inaccurate identification of anomalous data samples by an anomaly detection model, data drift can be identified, and the anomaly identification model can be updated to take this identified data drift into account when predicting future anomalous data samples.”)
Regarding claim 3, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
determining a distribution drift of the input data a difference between the anomaly scores of an earlier incoming data batch and the anomaly scores of a later incoming data batch is greater than a second threshold; and ([0049] “FIG. 3 depicts a series of stages 300-308 of calibration and operation of the components depicted in FIG. 2. In a first stage 300, the sensor device 118 initiates output of a stream of the frequency domain data 124 that is partitioned into a first time series 310 of frequency domain data followed by a second time series 318 of the frequency domain data. The first time series 310 represents normal operation of the asset (e.g., the rotational pump 106) and is used as training data to perform calibration of one or more machine-learning models and to generate reference residual data and anomaly score data, and the second time series 318 is used to monitor operation of the asset using the calibrated models.”)
providing a report relating to the determined distribution drift to a user, the respective device, an IT system connected to the respective device, or any combination thereof when if the determined difference is greater than a threshold. ([0035] “The GUI module 152 is executable by the one or more processors 120 to generate a graphical user interface 130 to display an alert indication 148. For example, the GUI module 152 may be executed by the one or more processors 120 to display the GUI 130 at the display device 128 to provide the operator 132 with the alert indication 148.”)
Regarding claim 4, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
assigning training data batches to the at least N separable classes of the anomaly detection models; and ([0049] “The first time series 310 represents normal operation of the asset (e.g., the rotational pump 106) and is used as training data to perform calibration of one or more machine-learning models and to generate reference residual data and anomaly score data”)
determining the given anomaly scores of the at least N separable classes for the N anomaly detection models. ([0052] “In a fourth stage 306, reference anomaly scores 242 are generated based on the reference residual data 236 and residual scores associated with a second portion 314 of the first time series 310.”)
Regarding claim 5, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
N=1. ([0028] “Thus, the described systems and methods address a significant challenge in deploying machine learning-based models at scale (e.g., individual models for a large number of assets). Techniques described herein, such as using models that are specific to a particular sensor pack and that primarily differentiate between normal and abnormal operation, enable a unique anomaly detection model to be created for an asset using a specific, common sensor pack without any human intervention. As a result, the described systems and methods can provide cost-beneficial machine-learning models for relatively large numbers of lower-cost assets that are not identical, such as pumps at an industrial plant.” This implies it could be one or more models).
Regarding claim 6, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
when the determined difference is smaller than the difference threshold: embedding the N anomaly detection models in a software application for analyzing, monitoring, operating, controlling, or any combination thereof of the at least one device; and ([0082] “The systems and methods illustrated herein may be described in terms of functional block components, screen shots, optional selections and various processing steps. It should be appreciated that such functional blocks may be realized by any number of hardware and/or software components configured to perform the specified functions.”).
deploying the software application on the at least one device or an IT system connected to the at least one device, such that the software application is usable for analyzing, monitoring, operating, controlling, or any combination thereof of the at least one device. ([0041] “Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
Regarding claim 7, Ardel in view of Schierz teaches claim 6 as outlined above. Schierz further teaches:
when the determined difference is greater than the difference threshold: amending the respective anomaly detection models, such that a determined difference using the respective amended anomaly detection models is smaller than the difference threshold; replacing the respective anomaly detection models with the respective amended anomaly detection models in the software application; and deploying the amended software application on the at least one device or the IT system. ([00198] “Responsive to an absolute value of the quantity difference being greater than a threshold difference, one or more actions associated with updating an anomaly detection model from which the first and second anomaly scores were obtained are performed. In some embodiments, the one or more actions can include determining that the features and/or the values of the features identified as contributing to the identification of data samples as anomalous in step 808 have changed between the first and second time points. In additional embodiments, the one or more actions can include providing a message to a user of the anomaly detection model, recommending that the anomaly detection model be updated. In additional embodiments, the one or more actions can include generating a new anomaly detection model based on the second plurality of data samples associated with the second time point. By generating the new anomaly detection model based on the second plurality of data samples, the new anomaly detection model can learn updated features that contribute to the identification of data samples as anomalous. By updating the anomaly detection model, the anomaly detection model can reduce (or avoid) flagging false positive anomalies or mistakenly identifying data samples as non-anomalous.”)
Regarding claim 9, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
for a plurality of interconnected devices: embedding respective N detection models in a respective software application for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected devices; ([0041] “Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
deploying the respective software application on the respective interconnected devices or an IT system connected to the plurality of interconnected devices, such that the respective software application is usable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected devices; ([0041] “Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
determining a respective difference of the respective anomaly detection models; and when the respective, determined difference is greater than a respective difference threshold: providing an alarm relating to the determined difference and the respective interconnected devices for which the corresponding respective software application used for analyzing, monitoring, operating, controlling, or any combination thereof of the respective interconnected device(s) devices to a user, the respective device, an automation system, or any combination thereof. ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
Regarding claim 10, Ardel in view of Schierz teaches claim 1 as outlined above. Ardel further teaches:
the respective device is a production machine, an automation device, a sensor, a production monitoring device, a vehicle or any combination thereof. ([0022] “ In some implementations, one such model uses data from one or more identical sensor packages which can be installed on a large number of similar or non-similar assets, such as rotational pumps that are deployed in relatively large numbers.”)
Regarding claim 11,
A system comprising: ([0024] “In illustrative implementation, a system includes a pre-processing module for data cleaning, a model generation model, and a post-processing module.”)
a first interface configured to receive input data relating to at least one device, wherein the input data comprises incoming data batches relating to at least N separable classes, with nϵ1, . . . , N; a computation unit configured to: ([0010] “In some aspects, a self-training sensor apparatus includes a sensor to generate sensor data associated with operation of a device and an input interface to receive input to select between a training mode of operation and a detection mode of operation. The apparatus also includes a memory including model generation instructions and one or more processors coupled to the memory. The one or more processors are configured to, while operating in the training mode of operation, receive first sensor data from the sensor and execute the model generation instructions based on the first sensor data to generate a trained anomaly detection model and an alert detection model. The one or more processors are also configured to while operating in the detection mode of operation, receive second sensor data from the sensor, use the trained anomaly detection model to generate an anomaly score, and process the anomaly score using the alert detection model to determine whether to generate an alert. The apparatus further includes an output interface to output an alert indication responsive to the one or more processors generating the alert.”)
determine respective anomaly scores for the respective incoming data batch relating to the at least N separable classes using N anomaly detection models; ([0034] “The one or more processors 120 are configured to process the frequency domain data 124 using the trained anomaly detection model 140, and the trained anomaly detection model 140 is configured to generate an anomaly score 142 for each sensing period.”)
determine, for the respective incoming data batch, a difference between the determined respective anomaly scores for the at least N separable classes and given respective anomaly scores of the N anomaly detection models; ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
and a second interface, configured to provide an alarm relating to the determined difference to a user, the respective device, an IT system connected to the respective device, or a combination thereof when the respective determined difference between is greater than a difference threshold. ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
wherein the at least N separable classes relate to typical scenarios of the respective device and correspond to a state of proper operation and to N-1 typical failure modes of the respective device, ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
wherein the N anomaly detection models include N trained anomaly detection models, the N trained anomaly detection models being trained using a reference data set that is provided beforehand by identifying the typical scenarios and related to typical variables or input data, and ([0005] “In some implementations, the models can be automatically generated and trained for each particular asset based on historic non-anomalous data for that particular asset.”)
wherein the typical scenarios include a scenario when the respective device is working properly, and when the respective device is damaged. ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
Ardel does not teach generate output data the generation of the output data comprising application of the anomaly detection models to the input data, the output data being suitable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective device; But Schierz does ([00104] “In some embodiments, an anomaly detection model simply provides a binary prediction of anomalousness of a data sample. For example, in some embodiments, an anomaly detection model can output a prediction of a data sample as anomalous or non-anomalous. In alternative embodiments, an anomaly detection model can generate an anomaly score for a data sample. An anomaly score for a data sample may indicate an extent to which the data sample is anomalous. In certain embodiments the extent to which a data sample is anomalous can be a probability that the data sample is anomalous and/or a degree to which the data sample is anomalous. For example, an anomaly detection model may generate an anomaly score of 0.9 for a data sample, thereby predicting that there is a 90% chance that the data sample is anomalous.”).
Ardel and Schierz are considered analogous art to the claimed invention because they are in the
same field of endeavor being anomaly detection. It would have been obvious to one of
ordinary skill in the art before the effective filing date of the claimed invention to combine the anomaly detection system/method of Ardel with the anomaly detection method of Schierz. One would want to do this to achieve more accurate anomaly detection.
Regarding claim 13,
In a non-transitory computer-readable storage medium that stores instructions executable by a system, the instructions comprising: ([0008] “In some aspect, a computer-readable storage device stores instructions”)
receiving input data relating to at least one device, wherein the input data comprise comprises incoming data batches relating to at least N separable classes, with nϵ1, . . . , N; ([0010] “In some aspects, a self-training sensor apparatus includes a sensor to generate sensor data associated with operation of a device and an input interface to receive input to select between a training mode of operation and a detection mode of operation. The apparatus also includes a memory including model generation instructions and one or more processors coupled to the memory. The one or more processors are configured to, while operating in the training mode of operation, receive first sensor data from the sensor and execute the model generation instructions based on the first sensor data to generate a trained anomaly detection model and an alert detection model. The one or more processors are also configured to while operating in the detection mode of operation, receive second sensor data from the sensor, use the trained anomaly detection model to generate an anomaly score, and process the anomaly score using the alert detection model to determine whether to generate an alert. The apparatus further includes an output interface to output an alert indication responsive to the one or more processors generating the alert.”)
determining respective anomaly scores for the respective incoming data batch relating to the at least N separable classes using N anomaly detection models; ([0034] “The one or more processors 120 are configured to process the frequency domain data 124 using the trained anomaly detection model 140, and the trained anomaly detection model 140 is configured to generate an anomaly score 142 for each sensing period.”)
determining, for the respective incoming data batch, a difference between the determined respective anomaly scores for the at least N separable classes and given respective anomaly scores of the N anomaly detection models; and ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
when the respective determined difference between is greater than a difference threshold, providing an alarm relating to the determined difference to a user, the respective device, an IT system connected to the respective device, or any combination thereof. ([0053] “The SPRT 208 uses the reference anomaly scores 242 (e.g., the known distribution of T.sup.2 scores from training) and the statistical data 240 corresponding to the anomaly scores 238 to determine whether to issue an alert based on the likelihood that the statistical data 240 represents a shift from the reference anomaly scores 242.”)
wherein the at least N separable classes relate to typical scenarios of the respective device and correspond to a state of proper operation and to N-1 typical failure modes of the respective device, ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
wherein the N anomaly detection models include N trained anomaly detection models, the N trained anomaly detection models being trained using a reference data set that is provided beforehand by identifying the typical scenarios and related to typical variables or input data, and ([0005] “In some implementations, the models can be automatically generated and trained for each particular asset based on historic non-anomalous data for that particular asset.”)
wherein the typical scenarios include a scenario when the respective device is working properly, and when the respective device is damaged. ([0041] “More generally, the system 100 enables detection of deviation from an operating state of the asset, such as detecting a transition from a first distinct operating state (e.g., the “normal” state to which the model is trained) to a second distinct operating state (e.g., the “abnormal” state). The second operating state, although distinct from the first operating state, may also be a “normal” operating state that is not associated with a malfunction or fault of the asset. Multiple instances of the sensor devices 118 and the alert detection device 110 can be deployed at low cost to process the frequency domain data 124 for a large number of assets, such as via wireless transmissions from the sensor devices 118 to a central server facility or cloud-based computing service. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
Ardel does not teach generating output data, the generating of the output data comprising applying the N anomaly detection models to the input data, the output data being suitable for analyzing, monitoring, operating, controlling, or any combination thereof of the respective device. But Schierz does ([00104] “In some embodiments, an anomaly detection model simply provides a binary prediction of anomalousness of a data sample. For example, in some embodiments, an anomaly detection model can output a prediction of a data sample as anomalous or non-anomalous. In alternative embodiments, an anomaly detection model can generate an anomaly score for a data sample. An anomaly score for a data sample may indicate an extent to which the data sample is anomalous. In certain embodiments the extent to which a data sample is anomalous can be a probability that the data sample is anomalous and/or a degree to which the data sample is anomalous. For example, an anomaly detection model may generate an anomaly score of 0.9 for a data sample, thereby predicting that there is a 90% chance that the data sample is anomalous.”).
Ardel and Schierz are considered analogous art to the claimed invention because they are in the
same field of endeavor being anomaly detection. It would have been obvious to one of
ordinary skill in the art before the effective filing date of the claimed invention to combine the anomaly detection system/method of Ardel with the anomaly detection method of Schierz. One would want to do this to achieve more accurate anomaly detection.
Regarding claim 14, Ardel in view of Schierz teaches claim 13 as outlined above. Ardel further teaches:
the input data undergoes a distribution drift involving an increase of the determined difference. ([00192] “To avoid inaccurate identification of anomalous data samples by an anomaly detection model, data drift can be identified, and the anomaly identification model can be updated to take this identified data drift into account when predicting future anomalous data samples.”)
Regarding claim 15, Ardel in view of Schierz teaches claim 13 as outlined above. Ardel further teaches:
determining a distribution drift of the input data a difference between the anomaly scores of an earlier incoming data batch and the anomaly scores of a later incoming data batch is greater than a second threshold; and ([0049] “FIG. 3 depicts a series of stages 300-308 of calibration and operation of the components depicted in FIG. 2. In a first stage 300, the sensor device 118 initiates output of a stream of the frequency domain data 124 that is partitioned into a first time series 310 of frequency domain data followed by a second time series 318 of the frequency domain data. The first time series 310 represents normal operation of the asset (e.g., the rotational pump 106) and is used as training data to perform calibration of one or more machine-learning models and to generate reference residual data and anomaly score data, and the second time series 318 is used to monitor operation of the asset using the calibrated models.”)
providing a report relating to the determined distribution drift to a user, the respective device, an IT system connected to the respective device, or any combination thereof when if the determined difference is greater than a threshold. ([0035] “The GUI module 152 is executable by the one or more processors 120 to generate a graphical user interface 130 to display an alert indication 148. For example, the GUI module 152 may be executed by the one or more processors 120 to display the GUI 130 at the display device 128 to provide the operator 132 with the alert indication 148.”)
Regarding claim 16, Ardel in view of Schierz teaches claim 13 as outlined above. Ardel further teaches:
assigning training data batches to the at least N separable classes of the anomaly detection models; and ([0049] “The first time series 310 represents normal operation of the asset (e.g., the rotational pump 106) and is used as training data to perform calibration of one or more machine-learning models and to generate reference residual data and anomaly score data”)
determining the given anomaly scores of the at least N separable classes for the N anomaly detection models. ([0052] “In a fourth stage 306, reference anomaly scores 242 are generated based on the reference residual data 236 and residual scores associated with a second portion 314 of the first time series 310.”)
Regarding claim 17, Ardel in view of Schierz teaches claim 13 as outlined above. Ardel further teaches:
N=1. ([0028] “Thus, the described systems and methods address a significant challenge in deploying machine learning-based models at scale (e.g., individual models for a large number of assets). Techniques described herein, such as using models that are specific to a particular sensor pack and that primarily differentiate between normal and abnormal operation, enable a unique anomaly detection model to be created for an asset using a specific, common sensor pack without any human intervention. As a result, the described systems and methods can provide cost-beneficial machine-learning models for relatively large numbers of lower-cost assets that are not identical, such as pumps at an industrial plant.” This implies it could be one or more models).
Regarding claim 19, Ardel in view of Schierz teaches claim 11 as outlined above. Ardel further teaches:
the system is an IT system. ([0024] “In illustrative implementation, a system includes a pre-processing module for data cleaning, a model generation model, and a post-processing module.”)
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Ardel in view of Schierz and Buesser (US 11334671).
Regarding claim 8, Ardel in view of Schierz teaches claim 6 as outlined above. Ardel further teaches:
analyzing, monitoring, operating, controlling, or any combination thereof of the at least one device using the backup software application. ([0072] “Thus, the method 600 enables distribution of a machine-learning type model that can be automatically trained for an individual asset (e.g., the specific rotational device 104) to distinguish between normal and abnormal operation and to generate an alert in response to detection of abnormal operation. As a result, machine-learning monitoring of a large number of relatively inexpensive assets can be deployed in a cost-effective manner.”)
Neither Ardel or Schierz teaches: when the amendment of the anomaly detection models takes more time than a duration threshold: replacing the deployed software application with a backup software application; However Buesser does (Col 3 lines 50-67 and Col 4 lines 1-4 “One or more roll-back strategies may be applied to maintain/keep intermediate model states (e.g., model weights) before some or all of them (e.g., keep/save model M2, M4, and M6 instead of saving models M1, M2, M3, and/or M5 so as to save maintenance costs) and recovers from the model states if a training collapse is detected. Training a machine learning model “M” means that the state of model M may be changed through a loop. At the beginning, M is at state M1, and the next iteration is state M2 and may continue until a last iteration with Mn. It may be expected that the model becomes more accurate/better through each iteration. So model state Mn may be the most optimized/best to be used for prediction. Models M1, M2, . . . Mn−1 may be called intermediate. However, in practice, it may not be known exactly what is a needed number of iterations (i.e. “n”). For example, perhaps model M10 is ideal and most optimal and it may be desirable stop when “n” is equal to ten (“10”) but that is unknown the system continues to train for M11 and so on which may be worse. Therefore, it is desirable to maintain/keep some intermediate model states just in cases they are better than the last model.”)
Ardel, Schierz and Buesser are considered analogous art to the claimed invention because they are in the same field of endeavor being training machine learning models. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the anomaly detection system/method of Ardel with the anomaly detection method of Schierz with the rollback strategy of Buesser. One would want to do this to not have downtime when training a model.
ConclusionApplicant'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 DANIEL P GRUSZKA whose telephone number is (571)272-5259. The examiner can normally be reached M-F 9:00 AM - 6:00 PM ET.
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, Li Zhen can be reached at (571) 272-3768. 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.
/DANIEL GRUSZKA/Examiner, Art Unit 2121
/Li B. Zhen/Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights