My research is mostly on helping humans interact with machine learning models meaningfully. That involves interpretability, trust, debugging, feedback, etc.