Abstract
Machine learning is not just about mastering board games like AlphaGo! This talk explores the powerful connection between two popular fields of research: machine learning and game theory. First, we will see how machine learning can tackle complex problems involving multiple players, not just games in the traditional sense. Dynamic stochastic games are crucial in fields like economics, ecology and operations research. We will also discuss links with AI for science. But the connection goes both ways: Game theory helps us understand some machine learning methods themselves. For instance, we will explain how generative adversarial networks (GANs) and diffusion models can be interpreted as games for neural networks. Last, we will give a few tips for doing research at the undergraduate level.
Biography
Mathieu Laurière is an assistant professor of Mathematics and Data Science at NYU Shanghai, he is also affiliated with the Courant Institute and the Center for Data Science, New York University. He was a postdoctoral researcher at NYU Shanghai and Princeton University, and a visiting faculty researcher at Google Brain in Paris. He received his Ph.D. from University Paris 7. Professor Lauriere’s main research interest is in stochastic optimal control, machine learning, game theory, and numerical methods.