Equifax Inc.

Michael McBurnett

Matthew Turner

Yafei Zhang

Lewis Jordan

Terry Woodford

Abhinav Sinha

Rongrong Dong

Michael Reith

Nikhil Paradkar

Patricia Bassetti

Jeffery Dugger

Techniques for risk evaluation include receiving, from a requesting entity, a request for monitoring target entities specifying a first identifier associated with each target entity and target entity information. The system generates a second identifier and a third identifier for each target entity and stores a mapping of the second identifiers to the first identifiers and the third identifiers, preventing the second identifiers from being provided to the requesting entity. The system monitors a periodically updated data set and determines risk metrics for the target entities, comparing each risk metric to a threshold value to identify target entities whose risk data indicates an insider threat. The system generates a third identifier for the identified target entities and provides the third identifiers to the requesting entity. Responsive to a request for a corresponding first identifier, the system identifies and provides the first and third identifiers to the requesting entity.

Continuous and anonymous risk evaluation

Certain embodiments involve generating or optimizing a neural network for generating analytical or predictive outputs. The neural network can be generated using a relationship between various predictor variables and an outcome (e.g., a condition's presence or absence). The neural network can be used to determine a relationship between each of the predictor variables and a response variable. The neural network can be optimized by iteratively adjusting the neural network such that a monotonic relationship exists between each of the predictor variables and the response variable. The optimized neural network can be used both for accurately determining response variables using predictor variables and determining adverse action codes for the predictor variables, which indicate an effect or an amount of impact that a given predictor variable has on the response variable. The neural network can be used to generate adverse action codes upon which consumer behavior can be modified to improve the response variable score.

Optimizing neural networks for generating analytical or predictive outputs

Certain aspects involve building timing-prediction models for predicting timing of events that can impact one or more operations of machine-implemented environments. For instance, a computing system can generate program code executable by a host system for modifying host system operations based on the timing of a target event. The program code, when executed, can cause processing hardware to a compute set of probabilities for the target event by applying a set of trained timing-prediction models to predictor variable data. A time of the target event can be computed from the set of probabilities. To generate the program code, the computing system can build the set of timing-prediction models from training data. Building each timing-prediction model can include training the timing-prediction model to predict one or more target events for a different time bin within the training window. The computing system can generate and output program code implementing the models' functionality.

Training or using sets of explainable machine-learning modeling algorithms for predicting timing of events

In some aspects, a computing system can obfuscate sensitive data based on data aggregation. A sensitive database containing sensitive data records can be joined with a grouping database containing a group identifier. The joining can be performed through a linking key that links a sensitive data record with a grouping data record in the grouping database. A grouping identifier can thus be obtained for each of the sensitive data record. The sensitive data records can be aggregated into aggregation groups based on their respective values of the group identifier. Statistics are calculated for the sensitive attributes of the sensitive data records in each aggregation group and are included in the aggregated data as the obfuscated version of the sensitive data. The aggregated data can be utilized to serve data queries from entities authorized or unauthorized to access the sensitive data.

Data protection via aggregation-based obfuscation

Certain aspects involve optimizing neural networks or other models for assessing risks and generating explanatory data regarding predictor variables used in the model. In one example, a system identifies predictor variables. The system generates a neural network for determining a relationship between each predictor variable and a risk indicator. The system performs a factor analysis on the predictor variables to determine common factors. The system iteratively adjusts the neural network so that (i) a monotonic relationship exists between each common factor and the risk indicator and (ii) a respective variance inflation factor for each common factor is sufficiently low. Each variance inflation factor indicates multicollinearity among the common factors. The adjusted neural network can be used to generate explanatory indicating relationships between (i) changes in the risk indicator and (ii) changes in at least some common factors.

Optimizing automated modeling algorithms for risk assessment and generation of explanatory data

Certain embodiments involve generating or optimizing a neural network for risk assessment. The neural network can be generated using a relationship between various predictor variables and an outcome (e.g., a condition's presence or absence). The neural network can be used both for accurately determining risk indicators or other outputs using predictor variables and for determining adverse action codes explaining the predictor variables' effect or an amount of impact on the risk indicator.

Optimizing neural networks for risk assessment

Certain embodiments involve generating or optimizing a neural network for risk assessment. The neural network can be generated using a relationship between various predictor variables and an outcome (e.g., a condition's presence or absence). The neural network can be used to determine a relationship between each of the predictor variables and a risk indicator. The neural network can be optimized by iteratively adjusting the neural network such that a monotonic relationship exists between each of the predictor variables and the risk indicator. The optimized neural network can be used both for accurately determining risk indicators using predictor variables and determining adverse action codes for the predictor variables, which indicate an effect or an amount of impact that a given predictor variable has on the risk indicator. The neural network can be used to generate adverse action codes upon which consumer behavior can be modified to improve the risk indicator score.

Certain aspects involve optimizing neural networks or other models for assessing risks and generating explanatory data regarding predictor variables used in the model. In one example, a system identifies predictor variables compliant with certain monotonicity constraints. The system generates a neural network for determining a relationship between each predictor variable and a risk indicator. The system performs a factor analysis on the predictor variables to determine common factors. The system iteratively adjusts the neural network so that (i) a monotonic relationship exists between each common factor and the risk indicator and (ii) a respective variance inflation factor for each common factor is sufficiently low. Each variance inflation factor indicates multicollinearity among a subset of the predictor variables corresponding to a common factor. The adjusted neural network can be used to generate explanatory indicating relationships between (i) changes in the risk indicator and (ii) changes in at least some common factors.

Growing community of inventors

Cumming, GA, United States of America

Michael McBurnett