Reinforcement and Mean-Field Games in Algorithmic Trading - Sebastian Jaimungal

Explore reinforcement learning and mean-field games in algorithmic trading through this comprehensive lecture by Prof. Sebastian Jaimungal at the Alan Turing Institute. Delve into two key areas of research: reinforcement learning techniques for solving stochastic control problems without explicit assumptions, and mean-field games with differing beliefs for optimizing trading actions among heterogeneous agents. Gain insights into double deep Q-learning, reinforced deep Kalman filters, and their applications in algorithmic trading. Examine the challenges of modeling interacting agents with disagreements on real-world models and their impact on trading strategies. Learn about latent factors driving prices, permanent price impact, and the resulting large stochastic game. Discover how these advanced concepts are applied to real data and their implications for financial markets.

Intro
Overview
Data
Limit order book
Control problem
Optimal solution
Reinforcement learning
Graphical model representation
Reinforcement
Neural nets
Heat map
Net results
Kalman filters
Maximum likelihood estimator
Batch reinforcement learning
Simultaneous analogous analysis