Abstract: Recently, the popularity and successes of Generative Adversarial Network (GANs) in computer vision and image generation have attracted intense attention from the mathematical finance community. GANs have since then been applied to financial data generation and lately shown capable of computing solutions for high dimensional mean-field games.

Despite these empirical successes of GANs, GANs training, usually via the stochastic gradient approach, has been challenging, especially in terms of its stability and convergence. In this talk, we discuss the game structure of GANs in an SDE framework, and present our latest work on the control approach for the stability of GANs training.

Based on joint work with Haoyang Cao of Alan Turing Institute and Othmane Mounjid of UC Berkeley.

Bio: Dr. Xin Guo is the Coleman Fung Chair Professor in Financial Modeling at the University of California, Berkeley and the Director of the Risk Analytics & Data Analysis Research Lab.

Dr. Xin Guo's research lies broadly in the span of stochastic controls and games, mean-field games, machine learning, mathematical finance, and FinTech. She has co-authored more than 70 research publications and a book in Quantitative Trading which has been translated into Chinese and Japanese.

Dr. Xin Guo is co-editor-in-chief of Frontier in Mathematical Finance, and the associate editor of many leading journals including Mathematical Finance, Operations Research, Mathematics, and Financial Economics. She is also the co-founder and co-chair of Women in Financial Engineering.

Meeting Recording: https://ucsb.zoom.us/rec/share/8Psk815MmoFGRxLNL8yURIbMzmqn4NVgYSs8WEdBdODroTiEJk9JyONkHhCdErQ.3QhrI8maCopwtSgc

Access Passcode: y0UBFk8$