Optimal Transport for Distributionally Robust Optimization and Applications in Machine Learning

Explore the intersection of Optimal Transport (OT) and Distributionally Robust Optimization (DRO) in this comprehensive seminar. Delve into how OT efficiently morphs probability distributions and how DRO tackles worst-case risk minimization under distributional ambiguity. Discover the computational advantages of OT-based DRO models and their applications in handling distribution shifts and heterogeneous data sources. Learn about the speaker's research in risk analytics and optimization, and gain insights into the role of OT-based DRO in responsible artificial intelligence. Examine how DRO concepts can be applied to solve complex OT problems, providing a holistic view of these interconnected fields in machine learning and decision-making under uncertainty.