Time: September 5 - December  7, Tuesday/Thursday, 11:00 AM - 12:20 PM MT. CAB 357

Recordings: Available on on request 

Overview

Traditional reinforcement learning (RL) systems assume a single agent interacts with a Markov decision process (MDP). A multi-agent RL setting assumes there are multiple agents. But what happens when one or more of the agents in a system is actually a human? Agents could learn from humans, humans could learn from agents, or agents and humans could learn to accomplish team goals together. This course will provide you with the background and tools needed to conduct research in this emerging area of human-in-the-loop (HitL) AI. 

More details are in the course syllabus.

Course Slack : https://join.slack.com/t/f23cmput656/shared_invite/zt-22f9tcjea-g29hSPMbVgllrnsPMHPGFg

Contact the instructor (Matt Taylor): matthew.e.taylor@ualberta.ca  

IRL Lab's webpage: http://irll.ca

More info about Matt: https://DrMattTaylor.net/