Abstract: The innovation disclosed and claimed herein, in one aspect thereof, comprises systems and methods of validating models guided by machine learning algorithms. The innovation can begin by receiving a risk model for validation having multiple sets of data. A first data set is selected from as an input. Outputs are generated for validation. One output can be generating a second set of analysis results using a comparable algorithm to the risk model. Another output can be generating a second set of variables and transformations using a machine learning algorithm and an untransformed set of the selected variables to assess the set of selected transformations. Another output can be generating a third set of variables using one or more machine learning algorithms and an extended feature set of variables to assess the selected variables. The outputs are compared to the analysis results, coefficients, selected variables, and selected transformations. A report of the comparison is generated.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating a predictive contribution report for an attribute using machine learning techniques. An example method includes generating an entity score for an entity using a predictive analysis machine learning model. The method further includes, in an instance the entity score fails to satisfy a determination decision threshold, selecting a reference entity from a plurality of candidate reference entities and determining a plurality of per-candidate feature contribution scores using a predictive analysis machine learning model. The method further includes generating a predictive contribution report, where the predictive contribution report includes an indication that the entity does not satisfy the determination decision threshold, and an indication of one or more candidate features associated with largest contributions to the entity score.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating a predictive temporal feature impact report using a feature engineering machine with attention for time series (FEATS model). An example method includes receiving an entity input data object. The method further includes determining one or more attention head scores for each feature attention head included in the FEATS model based at least in part on one or more per-temporal feature time impact scores over each time window for each temporal feature set. The method further includes generating a predictive temporal feature impact report based at least in part on at least one of the one or more attention head scores for each attention head or the one or more per-temporal feature time impact scores for each temporal feature time point as determined in each attention head.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating time series. A time series simulator receives information corresponding to a request for time series. The information is formatted into input data by the time series simulator. The input data comprises at least one continuous condition. A generator network of the continuous condition generative adversarial network (CCGAN) generates the time series based directly on a value of the at least one continuous condition. The time series is provided such that the time series is at least one of (a) provided as input to an analysis pipeline or (b) received by a user computing device wherein a representation of at least a portion of the one or more time series is provided via an interactive user interface of the user computing device.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating a single-index model (SIM) tree. An example method includes receiving a data set and a maximum tree depth. The example method further includes screening a set of variables from the data set to form split variables. The method may include, while maximum tree depth has not been reached, (i) generating a fast SIM estimation for nodes of a tree level, (ii) for each node, selecting a split point and split variable based on the fast SIM estimation, (iii) based on the selected split points and split variables, generating nodes for a next tree level, each including a subset of data, and (iv) repeating steps (i), (ii), and (iii). The method may include fitting a SIM for each leaf node at maximum tree depth based on a subset of the data set represented by the leaf node.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for determining robustness information for an NLP model. Modification rules, such as replacement rules and/or insertion rules, are used to generate instances of modified test data based on instances of test data that comprise words and have a syntax and a semantic meaning. The instances of test data and modified test data are provided to the NLP model and the output of the NLP model is analyzed to determine output changing instances of modified test data, which are instances of modified test data yielded output from the NLP model that is different and/or not similar to the output yielded from the NLP model for the corresponding instance of test data. Robustness information for the NLP model is determined based at least in part on the output changing instances of modified test data.

Abstract: Disclosed is an example approach in which network and non-network features are used to train a predictive machine learning model that is implemented to predict financial crime and fraud. Graphical network features may be generated by applying financial entity risk vectors to a network model with representations of various types of networks. The network model may comprise transactional, non-social, and/or social networks, with edges corresponding to linkages that may be weighted according to various characteristics (such as frequency, amount, type, recency, etc.). The graphical network features may be fed to the predictive model to generate a likelihood and/or prediction with respect to a financial crime. A perceptible alert is generated on one or more computing devices if a financial crime is predicted or deemed sufficiently likely. The alert may identify a subset of the set of financial entities involved in the financial crime and present graphical representations of networks and linkages.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating a single-index model (SIM) tree. An example method includes receiving a data set and a maximum tree depth. The example method further includes screening a set of variables from the data set to form split variables. The method may include, while maximum tree depth has not been reached, (i) generating a fast SIM estimation for nodes of a tree level, (ii) for each node, selecting a split point and split variable based on the fast SIM estimation, (iii) based on the selected split points and split variables, generating nodes for a next tree level, each including a subset of data, and (iv) repeating steps (i), (ii), and (iii). The method may include fitting a SIM for each leaf node at maximum tree depth based on a subset of the data set represented by the leaf node.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for determining robustness information for an NLP model. Modification rules, such as replacement rules and/or insertion rules, are used to generate instances of modified test data based on instances of test data that comprise words and have a syntax and a semantic meaning. The instances of test data and modified test data are provided to the NLP model and the output of the NLP model is analyzed to determine output changing instances of modified test data, which are instances of modified test data yielded output from the NLP model that is different and/or not similar to the output yielded from the NLP model for the corresponding instance of test data. Robustness information for the NLP model is determined based at least in part on the output changing instances of modified test data. White and/or black box attacks may be performed.

Abstract: Disclosed is an example approach in which network and non-network features are used to train a predictive machine learning model that is implemented to predict financial crime and fraud. Graphical network features may be generated by applying financial entity risk vectors to a network model with representations of various types of networks. The network model may comprise transactional, non-social, and/or social networks, with edges corresponding to linkages that may be weighted according to various characteristics (such as frequency, amount, type, recency, etc.). The graphical network features may be fed to the predictive model to generate a likelihood and/or prediction with respect to a financial crime. A perceptible alert is generated on one or more computing devices if a financial crime is predicted or deemed sufficiently likely. The alert may identify a subset of the set of financial entities involved in the financial crime and present graphical representations of networks and linkages.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating time series. A time series simulator receives information corresponding to a request for time series. The information is formatted into input data by the time series simulator. The input data comprises at least one continuous condition. A generator network of the continuous condition generative adversarial network (CCGAN) generates the time series based directly on a value of the at least one continuous condition. The time series is provided such that the time series is at least one of (a) provided as input to an analysis pipeline or (b) received by a user computing device wherein a representation of at least a portion of the one or more time series is provided via an interactive user interface of the user computing device.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for pricing a callable instrument. A plurality of corresponding pairs of Brownian motion paths and index value paths are determined corresponding to a set of dates. A deep neural network (DNN) of a backward DNN solver is trained until a convergence requirement is satisfied by for each pair of corresponding Brownian motion path and index value path, using the backward DNN solver to determine by iterating in reverse time order from a final discounted option payoff to an initial value. A statistical measure of spread of a set of initial values is determined and parameters of the DNN are modified based on the statistical measures of spread. Pricing information is determined by the backward DNN solver and provided such that a representation thereof is provided via an interactive user interface of a user computing device.

Abstract: Computing systems and technical methods that transform data structures and pierce opacity difficulties associated with complex machine learning modules are disclosed. Advances include a framework and techniques that include: i) global diagnostics; ii) locally interpretable models LIME-SUP-R and LIME-SUP-D; and iii) explainable neural networks. Advances also include integrating LIME-SUP-R and LIME-SUP-D approaches that create a transformed data structure and replicated modeling over local and global effects and that yield high interpretability along with high accuracy of the replicated complex machine learning modules that make up a machine learning application.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for screening data instances based on a target text of a target corpus. A screening device analyzes a target corpus to generate at least two term dictionaries for the target corpus. The screening apparatus, based on a frequency of a term in the target corpus, determines a term weight for the term; for each data instance, determines term scores for the data instance and the target text based on the term weights; filters the data instances based on the term scores, to generate a short list of data instances; determines term similarity scores between each data instance of the short list and target text based on the term weights; and provides a data instance determined to likely correspond to the target text and an indication of the corresponding term similarity score(s). A term is a word or an n-gram.

Abstract: The innovation disclosed and claimed herein, in one aspect thereof, comprises systems and methods of validating models guided by machine learning algorithms. The innovation can begin by receiving a risk model for validation having multiple sets of data. A first data set is selected from as an input. Outputs are generated for validation. One output can be generating a second set of analysis results using a comparable algorithm to the risk model. Another output can be generating a second set of variables and transformations using a machine learning algorithm and an un-transformed set of the selected variables to assess the set of selected transformations. Another output can be generating a third set of variables using one or more machine learning algorithms and an extended feature set of variables to assess the selected variables. The outputs are compared to the analysis results, coefficients, selected variables, and selected transformations. A report of the comparison is generated.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for pricing a callable instrument. A plurality of corresponding pairs of Brownian motion paths and index value paths are determined corresponding to a set of dates. A deep neural network (DNN) of a backward DNN solver is trained until a convergence requirement is satisfied by for each pair of corresponding Brownian motion path and index value path, using the backward DNN solver to determine by iterating in reverse time order from a final discounted option payoff to an initial value. A statistical measure of spread of a set of initial values is determined and parameters of the DNN are modified based on the statistical measures of spread. Pricing information is determined by the backward DNN solver and provided such that a representation thereof is provided via an interactive user interface of a user computing device.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for generating time series. A time series simulator receives information corresponding to a request for time series. The information is formatted into input data by the time series simulator. The input data comprises at least one continuous condition. A generator network of the continuous condition generative adversarial network (CCGAN) generates the time series based directly on a value of the at least one continuous condition. The time series is provided such that the time series is at least one of (a) provided as input to an analysis pipeline or (b) received by a user computing device wherein a representation of at least a portion of the one or more time series is provided via an interactive user interface of the user computing device.

Abstract: Generating, modeling, and operating optimal scorecards for credit risk evaluations is provided to a financial institution. Customer data is aggregated from a set of customer accounts. A score is generated for each product offered by a financial institution, where each score contributes to a plurality of combinations of scores. An aggregated model is generated based on the aggregated customer data and the generated scores. An aggregated score is computed using the aggregated model. In aspects of the subject innovation, the systems and methods disclosed leverage data from several sources and to include internal competitive and external competitive data to provide a more focused view of the consumer.

Abstract: Computing systems and technical methods that transform data structures and pierce opacity difficulties associated with complex machine learning modules are disclosed. Advances include a framework and techniques that include: i) global diagnostics; ii) locally interpretable models LIME-SUP-R and LIME-SUP-D; and iii) explainable neural networks. Advances also include integrating LIME-SUP-R and LIME-SUP-D approaches that create a transformed data structure and replicated modeling over local and global effects and that yield high interpretability along with high accuracy of the replicated complex machine learning modules that make up a machine learning application.

Abstract: A data driven and forward looking risk and reward appetite methodology for consumer and small business is described. The methodology includes account level historical data collection for customers associated with accounts as part of a portfolio. The account level historical data is segmented into groups of customers with similar revenues and loss characteristics. Segmented data is decomposed into seasoning, vintage, and cycle effects. Statistical clusters are formed based upon the data and effects. A simulation is applied to the statistical clusters and prediction data is generated. A simulation strategy to forecast and simulate revenue and loss volatility is developed. Efficient frontier curves of risk (e.g., return volatility) and reward (e.g., expected return) are created for the current portfolio under various economic scenarios.