Major Oral Presentation/Mathematical Finance Seminar: "Reinforcement Learning for Market Making"

Abstract: Market making is a fundamental trading problem in which an agent profits and provides liquidity by continually offering to buy and sell a security. Like many trading problems, it has become increasingly automated and we need to handle more data and act on ever-shorter timescales. That’s where the reinforcement learning method may fit in. In this talk, I’m going to show an interesting research problem where we train a market-making agent using reinforcement learning in a limit order book stock market. I’ll cover some of the basics in reinforcement learning and specify its application in our market making problem. A thorough comparison to other approaches in the literature will also be presented.

Host: Jose Figueroa-Lopez