题　目：Decision-Focused Learning: When and Why Traditional Prediction Models Fail

嘉　宾：刘墨 助理教授 北卡罗来纳大学教堂山分校

主持人：曾智宇 助理教授 上海交通大学安泰经济与管理

时　间：2026年3月16日（周一）10:00-11:30

地　点：安泰楼A503室

内容简介：

Plugging predictions of unknown parameters into downstream optimization problems, often referred to as the “predict-then-optimize” paradigm, has long been a standard approach in decision-making under uncertainty. However, improved predictive accuracy does not, in general, translate into improved decision quality. This disconnect has motivated growing interest in decision-focused learning within the operations research community. This talk will review recent developments in decision-focused learning and highlight key methodological insights, with a particular focus on stochastic linear programming as the downstream decision-making problem. We will discuss why several widely used tools in traditional statistical learning are not well suited for decision-focused settings and must be rethought, including (i) data collection strategies driven purely by predictive uncertainty and (ii) distributional distance measures such as the Wasserstein distance and Kullback–Leibler (KL) divergence.

演讲人简介：

Mo Liu is an assistant professor in the Department of Statistics and Operations Research at UNC Chapel Hill. His research interests center on decision-focused learning, a methodology that designs and trains prediction models to account for decision-making in downstream optimization problems. These downstream problems are typically linear optimization problems with real-world applications in revenue management, such as product recommendation, assortment optimization, and inventory management. Mo Liu received his Ph.D. in Industrial Engineering and Operations Research from the University of California, Berkeley, in 2024, and his B.S. in Industrial Engineering from Tsinghua University in 2019.

欢迎广大师生参加！