Interpretable Machine Learning for Materials Discovery Interpretable machine learning can do more than predict material performance. A study of spin-driven thermoelectric materials shows how transparent models helped researchers identify a useful property relationship and guide laboratory synthesis. Read More →