Interacting with Explanations to Improve Deep Neural Networks in Scientific Research

XIL allows scientists to interactively revise deep learning models, improving trust and performance by correcting "Clever Hans" moments.