Explainable AI: Is it Actually Important? Model Transparency & User Adoption

In predictive modeling a common experience the last decade is choosing between higher performance “black box” models at the expense of model interpretability and risking poor user adoption. Even with newer techniques to make black boxes transparent, the conversation isn’t over. Come hear how UnityPoint is addressing complex issues in user adoption such as inferring causality or succumbing to other human cognitive biases when “opening the black box”.