Abstract: Explanation of an analytical result, is afforded to a user by a populating a template with the result of searching homogenous clusters. During a preliminary phase, configuration changes are asynchronously fetched from services of an analytic application, and then grouped into homogenous clusters. Then, during a synchronous phase, a request to explain a particular analytical result is received from the application. Based upon content of the explanation request, the clusters are traversed in order to create a final path. A template comprising an explanation note with blanks, is selected from a template store and then populated with data from the final path. The populated template and the final path are stored together as an outcome. The outcome is then processed according to a challenge function, with the resulting challenged outcome communicated back to the application and afforded to provide the user with an explanation of the analytical result.

Abstract: A framework provides a detailed explanation regarding specific aspects of a (complex) calculation produced by an application (e.g., an analytical application). An explainability engine receives a request for explanation of the calculation. The explainability engine traverses homogenous data clusters according to the request, in order to produce a final path. The final path is used to select and then populate a template comprising explanation note(s). The outcome (comprising the final path and the template) is processed with a ruleset according to a covariance (COV) function in order to provide a first intermediate outcome. The first intermediate result is then processed with a second input according to a correlation (COR) function to provide a second intermediate outcome. The second intermediate result is processed according to a challenge function to provide a challenged outcome, and feedback (e.g., reward or penalization) to the ruleset. The challenged outcome provides detailed explanation to the user.

Abstract: Methods, systems, and computer-readable storage media for receiving a first bank statement at a hybrid pipeline including a set of lookup tables and a deep learning (DL) model that can each be used to determine customer IDs from bank statements, providing a first key based on data associated with the first bank statement, and determining that the first key is included in a first lookup table of the set of lookup tables, and in response: identifying a first set of customer IDs from the first lookup table, the first set of customer IDs including one or more customer IDs, and outputting the first set of customer IDs to computer-executable software that matches the first bank statement to one or more electronic documents at least partially based on the first set of customer IDs.

Abstract: Methods, systems, and computer-readable storage media for receiving historical data, the historical data including variable vectors, each variable vector being assigned to a class, processing the historical data through encoders to provide feature vectors, each feature vector corresponding to a respective variable vector and being assigned to the class of the respective variable vector, generating a set of decision trees based on the feature vectors, each decision tree corresponding to a class in the set of classes, transforming each decision tree into a set of rules to provide sets of rules, each rule in a set of rules defining conditions to assign at least a portion of an electronic document to a respective class in the set of classes, and providing the sets of rules for execution in an enterprise system, the enterprise system classifying electronic documents to classes in the set of classes based on the sets of rules.

Abstract: Methods, systems, and computer-readable storage media for receiving, by a ML application executing within a cloud platform, a first inference request, the first inference request including first inference data, transmitting, by the ML application, the first inference data to the UAT system within the cloud platform, retrieving, by the UAT system, a first ML model in response to the inference request, the first ML model being in an inactive state, providing, by the UAT system, a first inference based on the first inference data using the first ML model, providing a first accuracy evaluation at least partially based on the first inference, and transitioning the first ML model from the inactive state to an active state, the first ML model being used for production in the active state.

Abstract: As discussed herein, multiple neural networks are each trained on time-series data from a different domain. Each of the trained neural networks is used to make a domain-specific prediction for each point in time. Thus, time-series prediction data is generated by each of the trained neural networks. The domain-specific time-series prediction data are combined into a vector and used to train a final model that predicts a value. By breaking down the problem of forecasting into domain-specific forecasting models and a forecasting model, accuracy is improved over traditional document-based forecasting and computational resources are saved over traditional neural network designs.

Abstract: Methods, systems, and computer-readable storage media for receiving historical data, the historical data including variable vectors, each variable vector being assigned to a class, processing the historical data through encoders to provide feature vectors, each feature vector corresponding to a respective variable vector and being assigned to the class of the respective variable vector, generating a set of decision trees based on the feature vectors, each decision tree corresponding to a class in the set of classes, transforming each decision tree into a set of rules to provide sets of rules, each rule in a set of rules defining conditions to assign at least a portion of an electronic document to a respective class in the set of classes, and providing the sets of rules for execution in an enterprise system, the enterprise system classifying electronic documents to classes in the set of classes based on the sets of rules.

Abstract: Methods, systems, and computer-readable storage media for receiving user input indicating a first data point representative of output of a machine learning (ML) model, calculating a source model value based on the first data point and a second data point, calculating anti-model sub-values based on the first data point and a set of data points, providing an anti-model value based on the source model value and the anti-model sub-values, and determining a reliability of the output of the ML model based on the anti-model value.

Abstract: Methods, systems, and computer-readable storage media for receiving a first bank statement at a hybrid pipeline including a set of lookup tables and a deep learning (DL) model that can each be used to determine customer IDs from bank statements, providing a first key based on data associated with the first bank statement, and determining that the first key is included in a first lookup table of the set of lookup tables, and in response: identifying a first set of customer IDs from the first lookup table, the first set of customer IDs including one or more customer IDs, and outputting the first set of customer IDs to computer-executable software that matches the first bank statement to one or more electronic documents at least partially based on the first set of customer IDs.