Peter Carr Seminar Series: Anran Hu and Yu Yu
4 pm
Anran Hu
Title
Optimization and Learning for Mean-Field Games via Occupation Measure
Abstract
Mean-field games (MFGs) and multi-agent reinforcement learning (MARL) have become essential frameworks for analyzing interactions in large-scale systems. This talk presents recent advancements at the intersection of MFGs and MARL. We begin with a new framework MF-OMO (Mean-Field Occupation Measure Optimization), which reformulates Nash equilibria for discrete-time MFGs as a single optimization problem over occupation measures, offering a fresh characterization that enables the use of standard optimization algorithms to identify multiple equilibria without relying on restrictive assumptions. We also extend these results to continuous-time finite state MFGs.
Building on the concept of occupation measures, we then introduce MF-OML (Mean-Field Occupation Measure Learning), the first fully polynomial online RL algorithm capable of finding approximate Nash equilibria in large population games beyond zero-sum and potential games. We establish regret bounds for the $N$-player games that can be approximated by MFGs under monotonicity conditions. Together, these advancements provide a comprehensive approach to characterizing and solving Nash equilibria in complex multi-agent environments.
Bio
Anran Hu is an assistant professor at Columbia University, IEOR. She works at the intersection of stochastic control, game theory, optimization and machine learning. Her primary research areas are mean-field games, continuous-time stochastic control and reinforcement learning. She is also interested in FinTech and applying machine learning and reinforcement learning to finance. Before coming to Columbia University, Hu was a Hooke Research Fellow in the Mathematical Institute at the University of Oxford. She completed her Ph.D. in Berkeley IEOR, advised by Prof. Xin Guo. Before coming to Berkeley, she obtained my B.S. degree from the School of Mathematical Sciences, Peking University.
5 pm
Yu Yu
Title
Gen AI Use Case in Asset Management Industry and First-Hand Lessons Building Copilot
Abstract
As Gen AI technology improves in its performance and versatility, industry practitioners are actively evaluating and building Gen AI solutions. In this talk, I will discuss the high-level use cases I observe in the asset management industry, the approach business leaders use to make buy vs build decisions, as well as first-hand experience developing a co-pilot use case in a scalable fashion.
Bio
Yu Yu is a data science director at BlackRock, leading a team of data scientists to develop AI solutions. Her team has been building AI solutions for Global Client Business, Marketing, Client Experiences, and beyond. She has led the development of Experimental Design and Analysis for USWA, Client Tiering, and Segmentation, Fund Flow Drivers, OCIO Leads Creation, Client Lifetime Value, etc. Most recently she had been leading Gen AI data science development on RFP Companion and ChatBLK chatbot.
Prior to joining BlackRock, Yu Yu was a Director of Data Science at Bank of New York Mellon, where she built AI solutions that can help improve business outcomes for the bank as well as for its clients by driving automation, reducing risk, and delivering actionable insights. Her project on liquidity forecasting won the 2020 Gartner Eye on Innovation Award in Americas Financial Services. Before BNYM, Yu also worked at Point72 and AIG, where she leveraged cutting-edge statistical modeling and ML techniques to solve a wide range of investment and business problems. Yu was a tenure-track professor of marketing at Georgia State University for nearly five years prior to her industry career.
Yu Yu grew up in China and received her bachelor’s degree in finance from NanKai University. She went on to earn a master’s in economics from Indiana University—Bloomington and a PhD in Quantitative Marketing from Cornell University Johnson School of Management.
Password: PCAHYY129