Online Learning and Bandit Problems with Applications to Revenue Management

Abstract


1. Multi-armed Bandit Framework

2. The Upper Confidence Bound (UCB) Method

3. Elementary Improvements for Both Regret Minimization and Pure Exploration Settings

4. The Law-of-Iterated-Logarithm UCB Algorithm

5. Linear Contextual Bandits

6. Generalized Linear Models and Generalized Linear Contextual Bandits

7. Dynamic Assortment Optimization and the Multilinear Logit Bandit Problem

8. (If Time Permits) Zeroth Order Convex Optimization

Speaker


Yuan Zhou, Indiana University at Bloomington

Dr. Yuan Zhou is currently an Assistant Professor at the Computer Science Department of Indiana University at Bloomington. He is also a Visiting Assistant Professor at Shanghai University of Finance and Economics and an Adjunct Assistant Professor at University of Illinois at Urbana-Champaign. Before joining Indiana University, Yuan was an Applied Mathematics Instructor at the Mathematics Department of Massachusetts Institute of Technology. Prior to MIT, Yuan was the recipient of the Simons Graduate Fellowship and obtained his Ph.D. in Computer Science at Carnegie Mellon University. He also ranked the 1st in the International Olympiad in Informatics and the 2nd in the World Finals of ACM International Collegiate Programming Contest.

Yuan’s research interests include stochastic and combinatorial optimizations and their applications to operations management and machine learning. He is also interested in and publishes on analysis of mathematical programming, approximation algorithms, and hardness of approximation.

Time


2018.7.11, 7.13, 7.22, 7.27, 7.29, 8.5

9:00am

Room


Room 104, School of Information Management & Engineering, Shanghai University of Finance & Economics

Time


2018.7.15, 8.3

9:00am

Room


Room 102, School of Information Management & Engineering, Shanghai University of Finance & Economics