Observational Causal Inference for Improving and Auditing ML in Healthcare

Explore observational causal inference techniques for enhancing and auditing machine learning models in healthcare applications through this informative conference talk. Delve into the latest research on attributing model performance changes to distribution shifts, safe policy learning under partial identifiability, and robust off-policy evaluation using human inputs. Gain insights from related papers and learn how causal inference approaches can improve the reliability and interpretability of ML models in healthcare settings.

Shalmali Josh | Observational Causal Inference for Improving and Auditing ML for Healthcare | CGSI23