报告摘要：

Reinforcement learning (RL) means to “learn” the optimal control policy for an unknown system based on sampling and trial-and-error.

This talks gives a short introduction to the formulation and application of RL, its connection with control, as well as popular RL methods.
 
 
报告人简介: 

Mengdi Wang is an assistant professor at Department of Operations Research and Financial Engineering, Princeton University. She received her PhD from Department of EECS, Massachusetts Institute of Technology in 2013. Mengdi Wang is interested in stochastic optimization that involves uncertainty or hidden information, with an emphasis on optimization in machine learning and dynamic programming.