Abstract: Systems and methods are provided for accessing a database of records to identify a set of records represented by one or more nodes in a graph model. A connection between a first node and a second node in the one or more nodes is monitored to determine an association between a first record, represented by the first node, and a second record, represented by the second node. The set of records may be partitioned into a plurality of groups. For at least a first group, including a first set of records, it may be determined whether two or more records in the first group are related. In response to determining that the two or more records in the first group are related, a first group identifier may be assigned to the two or more records.

Abstract: Computer-implemented machines, systems and methods for providing insights about misalignment in a latent space of a machine learning model. A method includes initializing a second weight matrix of a second artificial neural network based on a first weight matrix from a first artificial neural network. The method further includes applying transfer learning between the first artificial neural network and the second artificial neural network. The method further includes comparing the first latent space with the second latent space. The method further includes determining, responsive to the comparing, a first score indicating alignment of the first latent space and the second latent space. The method further includes determining, and responsive to the first score satisfying a threshold, an appropriateness of the machine learning model.

Abstract: A method, a system, and a computer program product for generating a refined synthetic data from one or more sources of data. One or more source data are received from one or more data sources. One or more encoded source data are generated from the one or more source data. A synthetic data is generated by decoding one or more encoded source data. One or more variables in the synthetic data are selected and one or more predetermined identifiability values and one or more predetermined anonymity values are associated with them. The generated synthetic data including the selected variables is decoded using associated one or more predetermined identifiability values and one or more predetermined anonymity values. The decoded synthetic data is outputted.

Abstract: A method, a system, and a computer program product for generating an interpretable set of features. One or more search parameters and one or more constraints on one or more search parameters for searching data received from one or more data sources are defined. The data received from one or more data sources is searched using the defined search parameters and constraints. One or more first features are extracted from the searched data. The first features are associated with one or more predictive score values. The searching is repeated in response to receiving a feedback data responsive to the extracted first features. One or more second features resulting from the repeated searching are generated.

Abstract: A system and method for analyzing coverage, bias and model explanations in large dimensional modeling data includes discretizing three or more variables of a dataset to generate a discretized phase space represented as a grid of a plurality of cells, the dataset comprising a plurality of records, each record of the plurality of records having a value and a unique identifier (ID). A grid transformation is applied to each record in the dataset to assign each record to a cell of the plurality of cells of the grid according to the grid transformation. A grid index is generated to reference each cell using a discretized feature vector. A grid storage for storing the records assigned to each cell of the grid is then created. The grid storage using the ID of each record as a reference to each record and the discretized feature vector as a key to each cell.

Abstract: Computer-implemented methods, systems and products for analytics and discovery of patterns or signals. The method includes a set of operations or steps, including collecting data from a plurality of data sources, the data having a plurality of associated data types, and filtering the collected data based on identifying viable data sources from which the data is collected. The method further includes prioritizing discovery objectives based on analyzing the filtering results, and enriching the filtered collected data from viable data sources according to the prioritized discovery objectives. The method further includes extracting one or more signals from the enriched data using one or more machine learning mechanisms in combination with qualified subject matter expertise input, and graphically displaying the extracted signals in a meaningful way to a human operator such that the human operator is enabled to understand importance of extracted signals.

Abstract: A data object from a data source is received by a distributed process in a data stream. The distributed process has a sequence of categories, each category containing one or more tasks that operate on the data object. The data object includes files that can be processed by the tasks. If the task is able to operate on the data object, then the data object is passed to the task. If the task is unable to operate on the data object, then the files in the data object are passed to a file staging area of the distributed process and stored in memory. The files in the file staging area are passed, in sequence, from the file staging area to the task that was unable to operate on the data object. The data object is outputted to a next category or data sink after being operated on by the task.

Abstract: A method, a system, and a computer program product for generating an interpretable set of features. One or more search parameters and one or more constraints on one or more search parameters for searching data received from one or more data sources are defined. The data received from one or more data sources is searched using the defined search parameters and constraints. One or more first features are extracted from the searched data. The first features are associated with one or more predictive score values. The searching is repeated in response to receiving a feedback data responsive to the extracted first features. One or more second features resulting from the repeated searching are generated.

Abstract: Computer-implemented machines, systems and methods for managing missing values in a dataset for a machine learning model. The method may comprise importing a dataset with missing values; computing data statistics and identifying the missing values; verifying the missing values; updating the missing values; imputing missing values; encoding reasons for why values are missing; combining imputed missing values and the encoded reasons; and recommending models and hyperparameters to handle special or missing values.

Abstract: Systems and methods for providing insights about a machine learning model are provided. The method includes, using training data to train the machine learning model to learn patterns to determine whether data associated with an event provides an indication that the event belongs to a certain class from among a plurality of classes, evaluating one or more features of the machine learning model to produce a data set pairing observed scores S and a set of predictive input variables Vi, and constructing at least one data-driven estimator based on an explanatory statistic, the estimator being represented in a computationally efficient form and packaged with the machine learning model and utilized to provide a definition of explainability for a score generated by the machine learning model.

Abstract: A diagnostic system for model governance is presented. The diagnostic system includes an auto-encoder to monitor model suitability for both supervised and unsupervised models. When applied to unsupervised models, the diagnostic system can provide a reliable indication on model degradation and recommendation on model rebuild. When applied to supervised models, the diagnostic system can determine the most appropriate model for the client based on a reconstruction error of a trained auto-encoder for each associated model. An auto-encoder can determine outliers among subpopulations of consumers, as well as support model go-live inspections.

Abstract: Systems, methods and computer program products for improving security of artificial intelligence systems. The system comprising processors for monitoring one or more transactions received by a machine learning decision model to determine a first score associated with a first transaction. The first transaction may be identified as likely adversarial, in response to the first score being lower than a certain score threshold and the first transaction having a low occurrence likelihood. A second score may be generated in association with the first transaction based on one or more adversarial latent features associated with the first transaction. At least one adversarial latent feature may be detected as being exploited by the first transaction, in response to determining that the second score falls above the certain score threshold. Accordingly, an abnormal volume of activations of adversarial latent features spanning across a plurality of transactions scored may be detected and blocked.

Abstract: A sensitivity index model for predicting the sensitivity of an entity to a potential future disruption can be trained using a process that includes dividing a population of entities for which data attributes are available into matched pairs in a first sub-population and a second sup-population based on matching propensity scores for the entities using supervised machine learning, modeling outcomes for the two sub-populations, using the resultant models to calculate expected performances of the entities under differing conditions, and generating the sensitivity index model using supervised learning techniques based on quantification of differences between the calculated expected performances for the entities.

Abstract: Data characterizing an individual is received. Thereafter, one or more variables are extracted from the data so that, using a predictive model populated with the extracted variables, a likelihood of the individual adhering to a treatment regimen can be determined. The predictive model is trained on historical treatment regimen adherence data empirically derived from a plurality of subjects. Subsequently, data characterizing the determined likelihood of adherence can be promoted.

Abstract: Computer-implemented decision management systems and methods are provided. The method comprises obtaining information associated with factors usable for making a decision from among a plurality of inter-related decisions represented by a plurality of corresponding nodes. The computing environment provides access to resources that store information about relationships among the plurality of nodes. A relationship may be presentable as an edge connecting at least two nodes from among the plurality of nodes. The strength of the relationship between the at least two nodes is measurable and definable based on associations between the inter-related decisions. A valued may be determined that provides a measure for the strength of the relationship between the at least two nodes based on the information associated with the factors and the information about the relationships among the plurality of nodes.

Abstract: Systems and methods for generating concise explanations of scored observations that strike good, and computationally efficient, trade-offs between rank-ordering performance and explainability of scored observations are disclosed. The systems and methods described herein for explaining scored observations are based on a framework of partial dependence functions (PDFs), multi-layered neural networks (MNNs), and Latent Explanations Neural Network Scoring (LENNS).

Abstract: Systems, methods, and techniques to efficiently analyze and navigate through decision logic using an execution graph are provided. The method includes executing decision logic in response to receiving a data file. The method further includes generating, in response to the executing, an execution graph. The execution graph includes a plurality of nodes corresponding to a plurality of decision entities of the decision logic. The method further includes displaying the execution graph on a user interface. The method further includes displaying, in response to receiving a selection of a node of the plurality of nodes, information associated with the selected node.

Abstract: Systems, methods and products for quantitative translation of design requirements into a machine learning framework for training a classification model. A plurality of auxiliary tasks associated with a plurality of auxiliary task models are specified. The plurality of auxiliary task models are concurrently trained on the auxiliary tasks to generate one or more latent features learned by the plurality of auxiliary task models. The one or more latent features may be transferred from the plurality of auxiliary task models to augment a latent feature space of a target task for the classification model. Contribution levels of the transferred one or more latent features are adjusted based on design requirements for the target task for the classification model. First and second contribution levels are specified for respective first and second sets of auxiliary task latent features being quantified and enforced.

Abstract: A system and method for constructing an improved computing model that preserves use rights for data utilized by the model. A first dataset is accessed to build a computing model. The first data set is subject to terminable usage rights provisions. A portion of the first dataset is sampled to generate a second dataset. Vectors present in the first dataset and the second dataset are discretized. In response to determine that the usage rights associated with the primary dataset have been terminated, a coverage depletion for the second dataset is computed based on the usage rights termination associated with the first dataset. An estimated mean time to coverage failure for the first model based on the depletion coverage is determined for the second dataset. One or more data points are removed from the first dataset due to the termination of usage rights.

Abstract: Computer-implemented systems and methods for efficiently searching large data volumes for one or more items with a definable degree of similarity. The systems and methods may include functionality directed to selecting at least one token from the one or more tokens in a target item, the token including an identifiable character string defining, fully or partially, at least one of a name, an address, an entity or other identifier associated with the target item; extracting a character from the identifiable character string after the character string is standardized to a known common version of the character string; responsive to a character distribution lookup, determining that the extracted character corresponds to a first shard from among a plurality of discrete shards; and grouping the item into the first shard, the character distribution lookup being adjustable overtime to provide for a balanced distribution of items across the plurality of discrete shards.