Reinforcement learning (RL) has become an intensely studied research area in recent years. The majority of research papers in this area focus on advanced deep RL algorithms and incremental improvements in solution quality on standard environments. These include ATARI games, 2-D gridworlds, and more recently simulated dynamical systems such as the ones available in MuJoCo. Considering the gap between RL literature and the requirements of real-world problems, the goal of this workshop is to introduce the participants to the thought process behind solving real-world problems, and to help them formulate and solve an end-to-end problem over the course of two days.

Schedule:

Note that apart from the initial talk, the working sessions are open format with the participants expected to discuss and code together with their team. Detailed schedule is as follows:

Friday, 10:00-11:30 - Talk by Dr. Harshad covering the goal of the workshop, tips & tricks, some examples.
Friday, 11:30-11:45 - Refreshments Break
Friday, 11:45-13:00 - Ideation session (all teams must finalize what environment they want to work on)
Friday, 13:00-14:00 - Lunch Break- Lunch will be provided
Friday, 14:00-16:00 - Coding session 1

Saturday, 10:00-13:00 - Coding session 2 and Presentation preparation with a Refreshments break in between
Saturday, 13:00-14:00 - Lunch Break- Lunch will be provided
Saturday, 14:00-16:00 - Presentation and Evaluation