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 .
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 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.
DETAILED ACTION
Examiner Note
With regards to claims 12 and 17, the claims are directed toward statutory subject matter. Specifically, claim 12 and 17 each recites a “computer-readable medium” which can be interpreted to be or include signals or carrier waves. However, after a review of the specification, US 20250307390 A1, para. 0402 specifically states “The terms computer-readable media or computer-readable storage media do not include signals and carrier waves.” Therefore, the claims have not been interpreted to be directed towards non-statutory subject matter.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e. an abstract idea) without significantly more.
Step 1: This part of the eligibility analysis evaluates whether the claim falls within any statutory category. See MPEP 2106.03. The claims recites a computer-implemented method, One or more computer-readable media and a system. These are directed to a series of steps or acts, a manufacture (see above Examiner Note) and machine, and falls within one of the statutory categories of invention. (Step 1: YES).
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Claim 1, 10 and 17 are directed to an abstract idea because the following claim limitations recite an abstract idea:
A method, manufacture and machine comprising :
Fine-tuning a machine-learning (ML) tool; (mental process/mathematical concept: a human-being updating a mathematical model based on observation);
subsequently determining a current fingerprint of the ML tool; (mental process/mathematical concept: a human-being extracting and summating numerical parameters into a unified identifier);
comparing the current fingerprint with a prior fingerprint of the ML tool; (mental process/mathematical concept: a human-being comparing two values to compute a distance measure);
in a first case having a distance measure between the current and prior fingerprints above a predetermined threshold: (d) outputting a notification indicating detection of a threat. (mental process: a human-being noticing a difference and writing a warning note, texting/emailing another human being, etc.);
Claims 1, 10 and 17 recites the following additional elements:
Wherein the method is “computer-implemented”;
Wherein the manufacture is and includes “One or more computer-readable media storing instructions which, when executed on one or more hardware processors, cause the one or more hardware processors”
Wherein the machine is a “system” that comprises a “one or more hardware processors with memory coupled thereto” and “computer-readable media storing instructions which, when executed by the one or more hardware processors, cause the one or more hardware processors to perform operations ;
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application.
The claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements are recited at a high level of generality and amount to merely using computers as a tool to implement the abstract idea. Thus the additional elements are considered mere instruction to apply the abstract idea See MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to integrate the abstract idea into a practical application.
Step 2B:
This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
One way to determine integration into a practical application is when the claimed invention improves the functioning of a computer or improves another technology or technical field. To evaluate an improvement to a computer or technical field, the specification must set forth an improvement in technology and the claim itself must reflect the disclosed improvement. See MPEP 2106.04(d)(1) and 2106.05(a).
Likewise to step 2A prong 2, the claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements are recited at a high level of generality and amount to merely using computers as a tool to implement the abstract idea. Thus the additional elements are considered mere instruction to apply the abstract idea See MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to amount to significantly more than the abstract idea itself, even when the additional elements are considered alone and in combination with the abstract idea. (Step 2B: NO).
Therefore, the claims are directed to an abstract idea without significantly more and are unpatentable.
Claim 3
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Claim 3 are directed to an abstract idea because the following claim limitations recite an abstract idea:
Inherits the abstract idea of the base claim.
Claims 3 recites the following additional elements:
wherein the ML tool is part of a microservice in a network of microservices configured as a copilot.
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application.
The claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements merely limits implement the abstract idea to a generic technological environment or field of use, for example performing the abstract idea in a software environment. An abstract idea does not become nonabstract by limiting the invention to a particular field of use or technological environment. See MPEP 2106.05(h) and MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to integrate the abstract idea into a practical application.
Step 2B:
This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
One way to determine integration into a practical application is when the claimed invention improves the functioning of a computer or improves another technology or technical field. To evaluate an improvement to a computer or technical field, the specification must set forth an improvement in technology and the claim itself must reflect the disclosed improvement. See MPEP 2106.04(d)(1) and 2106.05(a).
Likewise to step 2A prong 2, the claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements merely limits implement the abstract idea to a generic technological environment or field of use, for example performing the abstract idea in a software environment. An abstract idea does not become nonabstract by limiting the invention to a particular field of use or technological environment. See MPEP 2106.05(h) and MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to amount to significantly more than the abstract idea itself, even when the additional elements are considered alone and in combination with the abstract idea. (Step 2B: NO).
Therefore, the claims are directed to an abstract idea without significantly more and are unpatentable.
Claim 11
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Claim 11 are directed to an abstract idea because the following claim limitations recite an abstract idea:
Inherits the abstract idea of the base claim.
Claims 11 recites the following additional elements:
wherein data inputted to the ML tool at act (a) is based on data outputted by another ML tool.
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application.
The claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements is merely a data gathering step necessary to obtain the input to perform the abstract idea and is considered to be insignificant extra-solution activity. See MPEP 2106.05(g). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to integrate the abstract idea into a practical application.
Step 2B:
This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
One way to determine integration into a practical application is when the claimed invention improves the functioning of a computer or improves another technology or technical field. To evaluate an improvement to a computer or technical field, the specification must set forth an improvement in technology and the claim itself must reflect the disclosed improvement. See MPEP 2106.04(d)(1) and 2106.05(a).
Likewise to step 2A prong 2, the claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements is merely a data gathering step necessary to obtain the input to perform the abstract idea and is considered to be insignificant extra-solution activity. See MPEP 2106.05(g). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to amount to significantly more than the abstract idea itself, even when the additional elements are considered alone and in combination with the abstract idea. (Step 2B: NO).
Therefore, the claims are directed to an abstract idea without significantly more and are unpatentable.
Claim 15 and 16
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Claims 15 and 16 are directed to an abstract idea because the following claim limitations recite an abstract idea:
wherein the ML tool is a neural network ML tool... each of the parameters reflecting a weight of a corresponding [first-tier] edge joining two cells of the neural network ML tool (mathematical concept: a human-being defining a graphical model with variables/parameters representing weighted links between nodes);
The parameters of the ML tool are allocated among a plurality of clusters/first-tier clusters, each of the clusters belonging to exactly one among a first group of clusters and a second group of clusters (mental process/mathematical concept: a human-being categorizing and sorting numerical variables into distinct sets/groups);
Other of the parameters reflecting weights of corresponding first-tier edges joining two cells of distinct first-tier clusters; (mathematical concept: a human-identifying variables/parameters that link distinct sets together);
Calculating respective first and second composite weight data structures with high and low fidelity for each of the clusters (mathematical concept: a human-being performing statistical calculations on the grouped variables/parameters for each distinct set );
Constructing a second-tier graph comprising a plurality of second-tier vertices...and a plurality of second-tier edges (mathematical concept: a human-being drawing a graph on a piece of paper);
Assigning a weight to teach second-tier edge, based on a number and/or weights of the first-tier edges (mathematical concept: a human-being calculating new values based on previous values);
Constructing one or more second composite weight data structures for the second-tier graph (mathematical concept: a human-being performing statistical calculations based on the graph variables/parameters);
Combining the first and second composite weight data structures into the current fingerprint (mathematical concept: a human-being aggregating the calculated numbers into a final numerical value);
Claims 15 and 16 recites the following additional elements:
None beyond the generic computing components recited in the base claim.
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application.
The claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements are recited at a high level of generality and amount to merely using computers as a tool to implement the abstract idea. The addition of neural network architecture, graphs, cluster weighting and weight calculations merely constitutes the addition of further abstract mathematical concepts and data organization techniques which does not make the claim patent-eligible simply by adding more specific mathematical formulas, rules or algorithms. See MPEP 2106.04(a)(2). Thus the additional elements are considered mere instruction to apply the abstract idea See MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to integrate the abstract idea into a practical application.
Step 2B:
This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
One way to determine integration into a practical application is when the claimed invention improves the functioning of a computer or improves another technology or technical field. To evaluate an improvement to a computer or technical field, the specification must set forth an improvement in technology and the claim itself must reflect the disclosed improvement. See MPEP 2106.04(d)(1) and 2106.05(a).
Likewise to step 2A prong 2, the claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional elements are recited at a high level of generality and amount to merely using computers as a tool to implement the abstract idea. The addition of neural network architecture, graphs, cluster weighting and weight calculations merely constitutes the addition of further abstract mathematical concepts and data organization techniques which does not make the claim patent-eligible simply by adding more specific mathematical formulas, rules or algorithms. See MPEP 2106.04(a)(2). Thus the additional elements are considered mere instruction to apply the abstract idea See MPEP 2106.05(f). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to amount to significantly more than the abstract idea itself, even when the additional elements are considered alone and in combination with the abstract idea. (Step 2B: NO).
Therefore, the claims are directed to an abstract idea without significantly more and are unpatentable.
Claim 20
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Claim 20 is directed to an abstract idea because the following claim limitations recite an abstract idea:
Inherits the abstract idea of the base claim.
Claims 20 recites the following additional elements:
at least one action to warn recipients of output from the ML tool between the prior fingerprint and the current fingerprint.
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application.
The claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional element merely constitutes a warning step necessary to be performed based on an output and is considered to be insignificant extra-solution post-solution activity. See MPEP 2106.05(g). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to integrate the abstract idea into a practical application.
Step 2B:
This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
One way to determine integration into a practical application is when the claimed invention improves the functioning of a computer or improves another technology or technical field. To evaluate an improvement to a computer or technical field, the specification must set forth an improvement in technology and the claim itself must reflect the disclosed improvement. See MPEP 2106.04(d)(1) and 2106.05(a).
Likewise to step 2A prong 2, the claims fails to achieve a technical solution to a technical problem. Thus the claim fail to provide an improvement to the function of a computer or to a technology itself. The claim culminate with outputting a notification indicating detection of a threat. See MPEP 2106.04(d)(1) and 2106.05(a). The additional element merely constitutes a warning step necessary to be performed based on an output and is considered to be insignificant extra-solution post-solution activity. See MPEP 2106.05(g). Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).Therefore, the examiner must find that the claims fail to amount to significantly more than the abstract idea itself, even when the additional elements are considered alone and in combination with the abstract idea. (Step 2B: NO).
Therefore, the claims are directed to an abstract idea without significantly more and are unpatentable.
Claims 2, 4-10, 13 and 14
Regarding claims 2, 4-10, 13 and 14 the following claim limitations recites an abstract idea
performing additional iterations. (mathematical concept: a human-being repeating calculations.)
determining composite weight data structures and combining the composite weight data structures (mathematical concept: a human being organizing numbers and performing mathematical operations on the organized numbers.)
using baseline fingerprints represented as vectors or points in a multiple dimensional space to calculated a Cartesian distance; (mathematical concept: a human-being performing geometric measurements.)
performing diagnostic actions, identifying impacted portions of the ML tool, identifying impacted input data, assessing the impact of anomalous data and performing generic remediation actions. (mental process: a human being evaluating a problem, identifying its root cause and implementing a solution based on the root cause.)
Claims 2, 4-10, 13 and 14 recites the additional elements:
none
Step 2A, Prong 2 and Step 2B
Claims 2, 4-10, 13 and 14 fail to recite any new additional elements relative to base claims 1, 10 and 17. Thus, the analysis and findings for step 2A, prong 2 and step 2B incorporates the analysis and findings of claims 1, 10 and 17 however, the analysis and findings includes consideration of claims 1, 10 and 17 as a whole. Therefore, claims 2, 4-10, 13 and 14 are directed to an abstract idea without significantly more and is unpatentable.
Claims 18 and 19
With regards to claims 18 and 19 these claims technical remediation steps that alter the routing of data or memory state of the ML tool (as opposed to merely outputting notifications or reports) and are deemed eligible under 35 USC 101, thus no rejection has been made.
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 5, 15 and 16 are 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 pre-AIA the applicant regards as the invention.
Regarding claim 5, the limitation “over multiple iterations” renders the claim indefinite because the base claim only discloses a single iteration, thus multiple iterations lacks an antecedent basis. The examiner recommends changing the dependency of claim 7 to claim 2 which states performing additional iterations or including language that provides proper basis for multiple iterations
Furthermore regarding claim 5, the claim references a step (e), however the base claim only recites steps (a) – (d).
Regarding claim 15, the limitations “high fidelity” and “low fidelity” renders the claim indefinite because the claims fails to establish an objective baseline, standard or measurable boundary for what constitutes “high” vs “low” fidelity. While the specification appears to define these terms based on specific loss percentages, 0050, these metrics are not incorporated in the claim language which leaves the metes and bounds of the claim unclear and subjective.
Regarding claim 16, the preamble of the claims recites “The computer-implemented method of claim 12” however claim 12 is directed towards a manufacture, specifically “one or more computer-readable media .” Thus as written claim 16, directed to a method, improperly depends on the manufacture of claim 12. For the purpose of examination the examiner will presume claim 16 to be directed to “one or more computer-readable media.”
Furthermore, regarding claim 16, the limitation “based on a number and/or weights of the first-tier edges joining the first and second first-tier clusters” renders the claim indefinite. The claim previously introduce “a plurality of first-tier clusters” and “distinct first-tier clusters” it’s unclear as to what “the first and second first-tier clusters” are supposed to represent.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 2, 4-6, 10-13, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over US 20240202405 to Lang et al. (hereinafter “Lang”) in view of US 20220383174 to Beveridge et al. (hereinafter “Beveridge”)
Claim 1
Lang teaches a computer-implemented method, comprising:
(a) fine-tuning a machine-learning (ML) tool; [e.g. Lang; Para. 0025, 0120-0139, 0147, 0167 – Lang discloses a surrogate model (e.g. a machine-learning tool) being fine-tuned.]
(b) subsequently determining a current version of the ML tool; [e.g. Lang; Para. 0120-0139, 0147, 0167 – Lang discloses determining a current version of the surrogate model that has been modified.]
(c) comparing the current version with a prior version of the ML tool; and in a first case having a distance measure between the current and prior versions above a predetermined threshold: [e.g. Lang; Para. 0120-0139, 0147, 0167, 0277 – Lang discloses comparing a current version of the surrogate model that has been modified to a previous or baseline version. Lang further discloses defining boundaries for normal and abnormal behavior (e.g. predetermined thresholds)]
(d) outputting a notification indicating detection of a threat. [e.g. Lang; Para. 0122, 0281, 0356, 0366 – Lang discloses outputting notifications.]
While Lang teaches the computer-implemented method of claim 1 and teaches identifying difference between machine learning models Lang fails to explicitly teach identifying this difference based on a fingerprint of the machine learning model and comparing the a current and prior fingerprint to using a distance measure.
However, Beveridge teaches: generating fingerprint known as model indicators of machine learning models based on artefacts/layers including weights and biases. Beveridge further discloses that fingerprints may be matrices of values and that similarity may be determined using Euclidean distance and tolerance values that gives meaning to the concept of nearness and how similar fingerprints are. See Abstract, Col 3 Ln 40 – Col 8 Ln 54.
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Beveridge fingerprint/model indicator comparison for comparing Lang current, new, modified surrogate model in the invention as disclosed by Lang as the comparison techniques are more computationally efficient and require less storage due, in part, to the artefact characterization and comparison techniques provided herein as disclosed by Beveridge Col 3 Ln 12-15. In combination this would provide a fingerprint comparison based on tolerance being outside of the boundary for model comparison and trigger a response.
Claim 2:
Lang teaches the computer-implemented method of claim 1, further comprising: performing additional iterations of acts (a)-(d). [e.g. Lang; Para. 0027, 0136-0139 – Lang discloses continuous training and analysis.]
Claim 4:
Lang in view of Beveridge teaches the computer-implemented method of claim 1, wherein the determining the current fingerprint comprises: for each of multiple clusters of the parameters of the ML tool: determining a composite weight data structure; and combining the composite weight data structures of the multiple clusters into the fingerprint. [e.g. Beveridge; Abstract, Col 3 Ln 40 – Col 8 Ln 54 – Beveridge discloses collecting and grouping artefacts such as layers of weights, bias, etc., and discloses generating a model indicator (e.g. composite weight data structure) for each layer and each fingerprint can collectively form the model indicator which is in the form of a matrix.]
Claim 5:
Lang and Beveridge teaches the computer-implemented method of claim 1, wherein the prior fingerprint comprises: a baseline fingerprint used as the prior fingerprint over multiple iterations of acts (a)-(e); or a fingerprint immediately preceding the current fingerprint. [e.g. Lang; Para. 0120-0139, 0365 – Lang discloses creating baseline models used for analysis.] [e.g. Beveridge; Abstract, Col 3 Ln 40 – Col 8 Ln 54 – Beveridge discloses generating fingerprint of machine learning models based on artefacts that comprises values affecting the operation of the machine learning model and comparing fingerprints through a similarity analysis such as determining a Euclidean distance and determining whether to take an action based on a threshold.]
Claim 6:
Lang as modified by Beveridge teaches the computer-implemented method of claim 1, wherein the current and prior fingerprints specify respective vectors or points in a multiple dimensional space and the distance measure comprises: a measure of Cartesian distance between the respective vectors or points. [e.g. Beveridge; Abstract, Col 3 Ln 40 – Col 8 Ln 54 – Beveridge discloses generating fingerprint of machine learning models based on artefacts that comprises values affecting the operation of the machine learning model and comparing fingerprints through a similarity analysis such as determining a Euclidean distance. The fingerprints are represented as a matrix of values (e.g. vectors or points in a multiple dimension space).]
Claim 10:
Lang teaches the computer-implemented method of claim 1, further comprising, in the first case, performing one or more remediation actions. [e.g. Lang; Para. 0122]
Claim 11:
Lang teaches the computer-implemented method of claim 1, wherein data inputted to the ML tool at act (a) is based on data outputted by another ML tool. [e.g. Lang; Para. 0130, 0138, 0139, 0140 – Lang discloses the training of the surrogate model includes using the outputs of another model (e.g. data inputted is based on data outputted of another ML tool).]
Regarding claims 12, 13 and 17 they are manufacture and system claims essentially corresponding to the above recitations, and they are rejected, at least, for the same reasons.
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over US 20240202405 to Lang et al. (hereinafter “Lang”) in view of US 20220383174 to Beveridge et al. (hereinafter “Beveridge”) and further In view of US 20250202843 to Bagga et al. (hereinafter “Bagga”)
Claim 3
While Lang and Beveridge teaches the computer-implemented method of claim 1 however the combination fails to explicitly teach the ML tool is part of a microservice in a network of microservices configured as a copilot.
However, Bagga teaches: digital assistant/workers implemented within a microservice based architecture and request handler components where the handlers apply ML models suitable for processing a specific type of request provided in a microservice based architecture. (e.g. ML tool is part of a microservice in a network of microservices configured as a copilot). See Para. 0032, 0034, 0036, 0037.
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Bagga microservice based digital assistant environment in the invention as disclosed by Lang as the microservice architecture makes troubleshooting more productive it would further enhance the anomaly/drift reports, alerts and mitigation actions of Lang making it more productive as these components would be able to be acted on by a conversational digital assistant that processes request and return responses without requiring any change to Lang surrogate model analysis.
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over US 20240202405 to Lang et al. (hereinafter “Lang”) in view of US 20220383174 to Beveridge et al. (hereinafter “Beveridge”) and further in view of NPL “Layer-wised Model Aggregation for Personalized Federated Learning” to Ma et al. (hereinafter “Ma”)
Claim 14
While Lang and Beveridge teaches the one or more computer-readable media of claim 13 and teach combining composite weight structures into the finger/model indicator however the combination fails to explicitly teach combining by applying weight factors to respective ones of the composite weight data structures.
However, Ma teaches: weighted aggregation of neural network model information on a per layer basis and teaches that different layers of a neural network have different utilities and as a result assigning unique weights to each layer rather than applying identical weights to all layers. Ma additionally teaches using an aggregation weight matrix based on aggregation weight vectors of n-th layers and obtaining model parameters by weighted aggregation according to the weight matrix. See Para. Abstract, Page 10092, 10093, 10095.
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ma weight factors to the per-layer/per-artefact fingerprint/model indicator structure of Beveridge in the invention as disclosed by the combination in order to account for the different utilities of the different layers of the machine learning model so that changes in most important layers have a greater effect on the final model indicator than changes in less important layers.
Claims 7-9 and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over US 20240202405 to Lang et al. (hereinafter “Lang”) in view of US 20220383174 to Beveridge et al. (hereinafter “Beveridge”) and further In view of US 20250202843 to Bagga et al. (hereinafter “Bagga”)
Claim 7
While Lang and Beveridge teaches the computer-implemented method of claim 1 and Lang and Beveridge both teach providing diagnostic actions however the combination fails to explicitly teach at least one action to identify one or more portions of the ML tool which were most impacted by the anomalous data.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0025, 0116, 0117 – Ezrielev discloses identifying model structures and weight view and parameters implicated by poisoned training data (e.g. identify one or more portions of the ML tool which were most impacted by the anomalous data).]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination in order to provide post detection diagnosis and remediation of anomalous model changes caused by poisoned or malicious training data. This would allow the system to predictably identify the model portions most impacted by poisoned or anomalous data.
Claim 8
While Lang and Beveridge teaches the computer-implemented method of claim 1 and Lang and Beveridge both teach providing diagnostic actions however the combination fails to explicitly teach at least one action to identify one or more portions of data inputted to the ML tool, between the prior fingerprint and the current fingerprint, which contributed to the distance measure exceeding the predetermined threshold.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0019, 0115, 0116, 0123 – Ezrielev discloses snapshotting a model, updating the model with new training data and indemnifying that poisoned data was introduced during the update and affects later model. (e.g. identify one or more portions of data inputted to the ML tool which contributed to the distance measure exceeding the predetermined threshold)]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination in order to provide post detection diagnosis and remediation of anomalous model changes caused by poisoned or malicious training data. This would allow the system to predictably identify the changes caused by poisoned or anomalous data between model states.
Claim 9
While Lang and Beveridge teaches the computer-implemented method of claim 1 and Lang and Beveridge both teach providing diagnostic actions however the combination fails to explicitly teach at least one action to assess an impact of the anomalous data on the ML tool.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0015, 0022, 0044, 0145 – Ezrielev discloses determining remediation based on the level of impact of the system caused by the poisoned (e.g. anomalous) data.]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination in order to provide post detection diagnosis and remediation of anomalous model changes caused by poisoned or malicious training data. This would allow the system to predictably assess the impact caused by the poisoned or anomalous data.
Claim 18
While Lang and Beveridge teaches the system of claim 17 however the combination fails to explicitly teach at least one action to control one or more sources of data inputted to the ML tool between the prior fingerprint and the current fingerprint.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0055, 0082, 0084 – Ezrielev identifies poisoned inferences and notifying inference consumers of the poisoned inference, wherein the notification identifies a time period associated with the poisoned inference.]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination in order to provide post detection diagnosis and remediation of anomalous model changes caused by poisoned or malicious training data. This would allow the system to predictably identify the sources of poisoned or anomalous data. Furthermore, removing the known source and implementing a snapshot to reduce computing resources spent re-training AI models as a snapshot can revert back to a last known good AI model and requires less resource expenditure than re-training an AI model from scratch as disclosed by Ezrielev in Para. 0018
Claim 19
While Lang and Beveridge teaches the system of claim 17 however the combination fails to explicitly teach at least one action to restore at least part of the ML tool to an earlier snapshot.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0018, 0055, 0054, 0085 – Ezrielev discloses remediation action that restore machine learning models to earlier snapshot.]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination in order to reduce computing resources spent re-training AI models as a snapshot can revert back to a last known good AI model and requires less resource expenditure than re-training an AI model from scratch as disclosed by Ezrielev in Para. 0018
Claim 20
While Lang and Beveridge teaches the system of claim 17 and do teach generating notifications however the combination fails to explicitly teach at least one action to warn recipients of output from the ML tool between the prior fingerprint and the current fingerprint.
However, Ezrielev teaches: generating and obtaining snapshots representing changes to an AI model and the snapshots store the model structure, architecture, parameters, weights, etc. and using this information to identify poisoned training data, data source, affected model information and remediation of the anomalous model change. [e.g. Ezrielev; Abstract Para 0055, 0082, 0084 – Ezrielev identifies poisoned inferences and notifying inference consumers of the poisoned inference, wherein the notification identifies a time period associated with the poisoned inference.]
Therefore, it would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to include Ezrielev diagnostic analysis in the invention as disclosed by the combination as this would allow the system to predictably communicate precise points when poisoned or anomalous data enter the machine learning model. Furthermore, implementing a snapshot to reduce computing resources spent re-training AI models as a snapshot can revert back to a last known good AI model and requires less resource expenditure than re-training an AI model from scratch as disclosed by Ezrielev in Para. 0018
Allowable Subject Matter
Claims 15 and 16 would be allowable if rewritten to overcome the rejection(s) under 35 U.S.C. 101 rejection, set forth in this Office action and to include all of the limitations of the base claim and any intervening claims.
Claims 15 and 16 would be allowable if rewritten to overcome the rejection(s) under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), 2nd paragraph, set forth in this Office action and to include all of the limitations of the base claim and any intervening claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CHRISTOPHER C HARRIS whose telephone number is (571)270-7841. The examiner can normally be reached Monday through Friday between 8:00 AM to 4:00 PM CST.
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, Jeffrey L Nickerson can be reached on (469) 295-9235. 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.
/CHRISTOPHER C HARRIS/Primary Examiner, Art Unit 2432
Prosecution Insights