Researchers create a mathematical framework to evaluate explanations of machine-learning models and quantify how well people understand them.
Read More