Deep Learning for Control

IAP 2021

Course Description

An overview of current deep reinforcement learning methods, challenges, and open research topics. The course will be taught by current members of the Improbable AI Lab at CSAIL, with the goal of providing a “bootcamp” for those wishing to get up to speed on current work in Robotics and Deep RL. Weeks 1-2 will detail understanding intelligence via machine learning to deep reinforcement architectures and frameworks (including methods for learning from demonstrations and practical RL). Week 3 will focus on learning for robotics and designing for efficient deep learning infrastructures.

Course Format

This is an IAP course that will be a mix of virtual lectures and homeworks. The plan is to delve into practical aspects of different algorithmic topics related to deep learning for control and follow it up with a homework. Students are encouraged to use Piazza (sign-up link) to discuss the topics covered in the lectures.


6.036 or 6.S191 (or any other equivalent courses): Working knowledge of machine learning, deep learning, reinforcement learning and linear algebra is assumed. Experience with deep learning packages such as PyTorch / Tensorflow is assumed. Homeworks will be in python. This is NOT a basic course in RL or Deep Learning or AI.


This is a new class and there is no textbook. Please refer to the schedule.


This is a new class and there is no textbook. We will post relevant reading material.

Links and due dates for assignments will be on the schedule.

Submission of Assignments
We'll be using Gradescope for problem set submission and grading. Each problem set is weighted equally. The login code for this class will be posted on Piazza. Grading will rely on review of the submitted code and writeup. More details will be provided when assignments are released.


You can contact the course staff at: rlcamp-staff [AT] mit [DOT] edu