Causal Reasoning in Healthcare: ML Guidelines & Dataset Shifts

This lecture by the instructor covers the practical application of causal reasoning in healthcare AI, focusing on emerging guidelines and regulations for ML in healthcare, obstacles to clinical translation of ML, dataset shifts in medical imaging, and the importance of causal diagrams. Through case studies in histopathology and breast screening, the lecture explores the impact of bias and the challenges of counterfactual inference. The presentation concludes with a discussion on how causality can aid in multimodal learning and improve biomedical vision-language processing.