Wells touts new explainability technique for AI credit models

Novel interpretability method could spur greater use of ReLU neural networks for credit scoring