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: 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: Failed circuits (e.g., defects) on each of a plurality of semiconductor wafers (10) in a batch can be characterized for the purpose of identifying defect sources by first mapping the defective ones of the circuits (12.sub.1 -12.sub.n) in each wafer. A determination is made to see if the defects in the defect pattern map associated with each wafer (10) are sufficiently clustered to warrant further study. The defect pattern maps for the wafers in the batch identified as having spatial clustering present are smoothed and thresholded to identify where spatial clusters occur. All such smoothed and thresholded defect pattern maps are separated into groups in accordance with the pattern of defects. The pattern of defects associated with each group is then analyzed to determine if any relationship exists between the pattern and the order of the process steps or one of the patterns in a library of patterns associated with particular failure modes.