DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
This office action is responding to the amendments filed on 05/22/2026. Claims 1, 2 and 12 have been amended.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-2, 5-6, 8, 10, 12, 15-16, 18 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Yang et al. [US 2021/0374132 A1] in view of Cella [US 2019/0340716 A1] and in further view of Friedman et al. [US 2022/0165007 A1].
Claim 1 is rejected over Yang, Cella and Friedman.
Yang teaches “A system for using explainability vectors to rank user interface elements, the system comprising:” as “The system provides the recommendation for one or more items and corresponding explanation narratives based on ranking predicted scores and explainability scores for the items.” [Abstract]
“receiving training data for a predictive machine learning model that outputs resource availability scores, wherein the training data comprises values for a first set of features, wherein the first set of features comprises variables that influence resource availability;” as “recommendation explainability scores or metrics can provide explanations to users for why certain items are recommended.” [¶0016]
“training the predictive machine learning model based on the training data;” as “such processors may accelerate various computing tasks associated with evaluating neural network models (e.g., training, prediction, preprocessing, and/or the like) by an order of magnitude or more in comparison to a general-purpose CPU. ” [¶0025]
“processing the predictive machine learning model to extract an explainability vector, wherein each entry in the explainability vector corresponds to a feature in the first set of features and is indicative of a correlation between the feature and output of the predictive machine learning model;” as “relevance model 305 may be trained to learn the probability P(y|x) of an action label y given the input features x and generate a predicted score 320.” [¶0041]
“processing the second set of features and the output of the predictive machine learning model to generate an explanative factor;” as “ The system obtains first features of at least one user and second features of a set of items. The system provides the first features and the second features to a first machine learning network for determining a predicted score for an item.” [Abstract]
“training the ranking machine learning model, wherein the ranking machine learning model takes the third set of features and the explanative factor as input; and” as “Recommendation module 130 may then rank the combined score from the highest score to the lowest score and select an item that corresponds to the highest score or item(s) that corresponds to the top k scores as the recommended item(s).” [¶0075]
Yang does not explicitly teach processing the second set of features and the output of the predictive machine learning model to generate an explanative factor that indicates a task type causing one or more of particular resource consumption or impacting resource availability;
based on the explainability vector, processing the first set of features to generate a second set of features such that each feature in the second set of features has a correlation with the output of the predictive machine learning model that is above a correlation threshold;
determining to train a ranking machine learning model which uses a third set of features as input, wherein the third set of features contains variables affecting resource availability;
receiving, as output from the ranking machine learning model, a vector indicating display positions and rankings of one or more user interface elements for a software application.
However, Cella teaches “processing the second set of features and the output of the predictive machine learning model to generate an explanative factor that indicates a task type causing one or more of particular resource consumption or impacting resource availability;” as “With further reference to FIG. 57, the example system 5700 includes the controller 5702 configured to interpret a resource utilization requirement 5704 for a task system 5706, and to interpret a behavioral data source 5720; to operate a machine (e.g., an expert system 5710, and/or an AI or machine learning component) to forecast a forward market price 5712 for a resource in response to the resource utilization requirement 5704 and the behavioral data source 5720, and to perform one of adjusting an operation of the task system (e.g., providing operational adjustments 5722) or executing a transaction 5714 in response to the forecast of the forward market price 5712 for the resource.” [¶1075]
Yang and Cella are analogous arts because they teach machine learning and adaptive learning.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Yang and Cella before him/her, to modify the teachings of Yang to include the teachings of Cella with the motivation of a platform for enabling transactions is provided having a machine with a regenerative energy facility that optimizes allocation of delivery of energy produced among compute tasks, networking tasks and energy consumption tasks. [Cella, ¶0649]
The combination of Yang and Cella does not explicitly teach based on the explainability vector, processing the first set of features to generate a second set of features such that each feature in the second set of features has a correlation with the output of the predictive machine learning model that is above a correlation threshold;
determining to train a ranking machine learning model which uses a third set of features as input, wherein the third set of features contains variables affecting resource availability;
receiving, as output from the ranking machine learning model, a vector indicating display positions and rankings of one or more user interface elements for a software application.
However, Friedman teaches “based on the explainability vector, processing the first set of features to generate a second set of features such that each feature in the second set of features has a correlation with the output of the predictive machine learning model that is above a correlation threshold;” as “The computing machine computes, for each first node from among a plurality of first nodes that are intermediate nodes or end nodes, a provenance value representing dependency of an explainability value vector of the first node on the one or more nodes upstream from the first node. The computing machine computes, for each first node, the explainability value vector. The computing machine provides a graphical output representing at least an explainability value vector of an end node.” [Abstract]
“determining to train a ranking machine learning model which uses a third set of features as input, wherein the third set of features contains variables affecting resource availability;” as “the terms “first,” “second,” and “third,” etc. are used merely as labels, and are not intended to impose numerical requirements on their objects.” [¶0216]
“receiving, as output from the ranking machine learning model, a vector indicating display positions and rankings of one or more user interface elements for a software application.” as “using a rover's GPS sensor to measure its position assumes that the GPS sensor is on the rover. This assumption affects the integrity of all downstream beliefs and planned actions that rely directly or indirectly on positional data.” [¶0170]
Yang, Cella and Friedman are analogous arts because they teach machine learning and adaptive learning.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Yang, Cella and Friedman before him/her, to modify the teachings of combination of Yang and Cella to include the teachings of Friedman with the motivation of the provenance-based approach. treats the plan as a tripartite dependency graph that helps explain the foundations, reliability, impact, and sensitivity of the information that comprises the plan's states and actions. [Friedman, ¶0135]
Claim 2 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 1.
Claim 5 is rejected over Yang, Cella and Friedman.
Yang teaches “wherein processing the second set of features and the output of the predictive machine learning model to generate an explanative factor comprises: generating an encoding map which translates the first set of features to the second set of features;” as “support vector machines, encoders, decoders, auto-encoders, stacked auto-encoders, perceptrons, multi-layer perceptrons, artificial neural networks” [¶0030]
The combination of Yang and Cella does not explicitly teach using the output of the predictive machine learning model and the explainability vector, generating an embedding vector; and
based on the encoding map and the embedding vector, generating the explanative factor.
However, Friedman teaches “using the output of the predictive machine learning model and the explainability vector, generating an embedding vector; and” as “each beginning node in at least a subset of the beginning nodes having an explainability value vector. ” [¶0007]
“based on the encoding map and the embedding vector, generating the explanative factor.” as “The activity record can include an association with the received datum and any input datums used by the agent to generate the received datum.” [¶0052]
Claim 6 is rejected over Yang, Cella and Friedman.
Yang teaches “using the vector indicating rankings of one or more user interface elements, determining a display order of the one or more user interface elements; and based on the display order of the one or more user interface elements, causing to be displayed on a user interface on a user device the one or more user interface elements.” as “The data may also be segmented into training and test datasets based on user rating history in a leave-one-out way. For example, for each user, the movies the user rated may be sorted by the timestamp in ascending order. ” [¶0078]
Claim 8 is rejected over Yang, Cella and Friedman.
Yang teaches “the predictive machine learning model is defined by a set of parameters comprising a matrix of weights for a supervised classifier algorithm; and” as “Further a machine learning process may comprise a trained algorithm that is trained through supervised learning (e.g., various parameters are determined as weights or scaling factors).” [¶0030]
“the explainability vector is extracted from the set of parameters using a Local Interpretable Model-agnostic Explanations method.” as “Note that the LIME method is a model-agnostic method for generating explanations, requiring training of a local linear model for each user and item pair.” [¶0093]
Claim 10 is rejected over Yang, Cella and Friedman.
Yang teaches “the predictive machine learning model is defined by a set of parameters comprising a matrix of weights for a convolutional neural network algorithm; and” as “the module may be implemented on one or more neural networks, such as one or more supervised and/or unsupervised neural networks, convolutional neural networks, and/or memory-augmented neural networks, among others.” [¶0023]
“the explainability vector is extracted from the set of parameters using a Gradient Class Activation Mapping method.” as “The machine learning process may comprise one or more of regression analysis, regularization, classification, dimensionality reduction, ensemble learning, meta learning, association rule learning, cluster analysis, anomaly detection, deep learning, or ultra-deep learning. The machine learning process may comprise, but is not limited to: k-means, k-means clustering, k-nearest neighbors, learning vector quantization, linear regression, non-linear regression, least squares regression, partial least squares regression, logistic regression, stepwise regression, multivariate adaptive regression splines, ridge regression, principle component regression, least absolute shrinkage and selection operation, least angle regression, canonical correlation analysis, factor analysis, independent component analysis, linear discriminant analysis, multidimensional scaling, non-negative matrix factorization, principal components analysis, principal coordinates analysis, projection pursuit, Sammon mapping, t-distributed stochastic neighbor embedding, AdaBoosting, boosting, gradient boosting” [¶0030]
Claim 12 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 1.
Claim 15 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 5.
Claim 16 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 6.
Claim 18 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 8.
Claim 20 is rejected over Yang, Cella and Friedman under the same rationale of rejection of claim 10.
Claim(s) 7, 9, 17 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Yang et al. [US 2021/0374132 A1] in view of Cella [US 2019/0340716 A1] in further view of Friedman et al. [US 2022/0165007 A1] and in further view of Shah [US 2025/0029105].
Claim 7 is rejected over Yang, Cella, Friedman and Shah.
Yang teaches “the predictive machine learning model is defined by a set of parameters comprising a matrix of weights for a multivariate regression algorithm; and” as “The machine learning process may comprise one or more of regression analysis, regularization, classification, dimensionality reduction, ensemble learning, meta learning, association rule learning, cluster analysis, anomaly detection, deep learning, or ultra-deep learning. ” [¶0030]
The combination of Yang, Cella and Friedman does not explicitly teach the explainability vector is extracted from the set of parameters using a Shapley Additive Explanation method.
However, Shah teaches “the explainability vector is extracted from the set of parameters using a Shapley Additive Explanation method.” as “the fraud detection system 102 can determine a Shapley Additive Explanations (SHAP) value for each feature.” [¶0054]
Yang, Cella, Friedman and Shah are analogous arts because they teach machine learning and adaptive learning.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Yang, Cella, Friedman and Shah before him/her, to modify the teachings of combination of Yang, Cella and Friedman to include the teachings of Shah with the motivation of the fraud detection system, however, identifies features associated with the network transaction and utilizes a trained card-not-present machine learning model to generate accurate fraud predictions in real time. [Shah, ¶0022]
Claim 9 is rejected over Yang, Cella, Friedman and Shah.
The combination of Yang, Cella and Friedman does not explicitly teach the predictive machine learning model is defined by a set of parameters comprising a vector of coefficients for a generalized additive model; and
the explainability vector is extracted from the vector of coefficients in the generalized additive model.
However, Shah teaches “the predictive machine learning model is defined by a set of parameters comprising a vector of coefficients for a generalized additive model; and” as “the fraud detection system 102 can determine a Shapley Additive Explanations (SHAP) value for each feature.” [¶0054]
“the explainability vector is extracted from the vector of coefficients in the generalized additive model.” as “the card-not-present machine learning model 302 is a different type of machine learning model, such as a neural network, a support vector machine, or a random forest. ” [¶0058]
Yang, Cella, Friedman and Shah are analogous arts because they teach machine learning and adaptive learning.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Yang, Cella, Friedman and Shah before him/her, to modify the teachings of combination of Yang, Cella and Friedman to include the teachings of Shah with the motivation of the fraud detection system, however, identifies features associated with the network transaction and utilizes a trained card-not-present machine learning model to generate accurate fraud predictions in real time. [Shah, ¶0022]
Claim 17 is rejected over Yang, Cella, Friedman and Shah under the same rationale of rejection of claim 7.
Claim 19 is rejected over Yang, Cella, Friedman and Shah under the same rationale of rejection of claim 9.
Allowable Subject Matter
Claims 3-4, 11 and 13-14 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
The prior art of record, including Yang et al. [US 2021/0374132 A1], Cella [US 2019/0340716 A1], and Friedman et al. [US 2022/0165007 A1], whether considered alone or in combination, fails to teach or reasonably suggest the claimed specific processing of an explainability vector to generate a second set of features.
In particular, the prior art does not teach normalizing an explainability vector into a standard-deviation space to produce a processed vector, generating a covariance matrix based on the processed vector, computing eigenvectors for the covariance matrix, selecting a measure of coverage, selecting a subset of eigenvectors based on the measure of coverage, and determining the second set of features corresponding to the selected subset of eigenvectors. While the cited references may generally relate to machine learning, model interpretation, feature analysis, or explainability, they do not disclose the claimed ordered dimensionality-reduction/extraction pipeline applied specifically to an explainability vector.
The prior art also fails to teach or suggest generating a correlation matrix based on the explainability vector, computing eigenvectors for the correlation matrix, determining a threshold value using a distribution of the eigenvectors, and using a maximum-likelihood estimator model that receives the threshold value as an input to extract the second set of features from the correlation matrix. The claimed use of an eigenvector-distribution-derived threshold as an input to a maximum-likelihood estimator model for extracting features from an explainability-vector-based correlation matrix is not shown or suggested by the cited combination.
Furthermore, the cited references fail to teach or suggest a predictive machine learning model defined by parameters including a hyperplane matrix for a support vector machine algorithm, wherein the explainability vector is extracted from the model parameters using a counterfactual explanation method. Although counterfactual explanations and support vector machine models are known generally, the prior art does not disclose the claimed relationship in which the explainability vector is extracted from a set of SVM parameters including the hyperplane matrix using a counterfactual explanation method.
Accordingly, the claimed subject matters in claim 3-4, 11 and 13-14 are considered novel because the cited references, individually or in combination, do not teach or render obvious the specific claimed feature-generation techniques based on covariance/eigenvector coverage, correlation/eigenvector-threshold maximum-likelihood estimation, and SVM-parameter-based counterfactual extraction of an explainability vector.
Response to Arguments
Applicant’s arguments with respect to claim(s) have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
Applicant'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 MASUD K KHAN whose telephone number is (571)270-0606. The examiner can normally be reached Monday-Friday (8am-5pm).
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, Hosain Alam can be reached at (571) 272-3978. 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.
/MASUD K KHAN/Primary Examiner, Art Unit 2132
Prosecution Insights