PC5 Deep Reinforcement Learning

Description

Deep Reinforcement Learning (RL) is at the forefront of Artificial Intelligence research. In this Practical Course, you’ll understand the most recent advances in Deep RL, learn cutting-edge algorithms like Deep Q-Learning (DQN), Deep Deterministic Policy Gradients (DDPG), Actor-Critic, etc., and train your own AI agents in simulated environments.

Please bring your laptops.

Objectives

You’ll train AI agents to walk, drive, manipulate objects, or perform other complex tasks in simulated environments [1, 2, 3, 4]. This course consists of demonstrations, crystal-clear examples of code implementation, and practical exercises which you will perform on your laptops.

[1] https://rebrand.ly/HandManipulateBlock
[2] https://rebrand.ly/NIPS17L2R
[3] https://www.pommerman.com
[4] https://youtu.be/Ul0Gilv5wvY

Literature

Deep Q-Network & Dueling network architectures for deep reinforcement learning
https://youtu.be/XjsY8-P4WHM?t=48s

Introduction to Reinforcement Learning
http://videolectures.net/deeplearning2016_pineau_reinforcement_learning

Spinning Up in Deep RL!
https://spinningup.openai.com/en/latest/index.html

Exercises and solutions for popular Reinforcement Learning algorithms
https://github.com/dennybritz/reinforcement-learning

Course location

Günne

Course requirements

- basic Python experience: you can write nested for loops, define functions, etc., and read and understand code written by others.
- basic knowledge of machine learning techniques: you have seen a few examples of neural network architecture (e.g., CNN), and can explain what means backpropagation (https://www.youtube.com/results?search_query=backpropagation).
- personal laptop: Linux, Mac and Windows machines are equally fine for the purpose of this course.

Instructor information.

Instructor
Andrew Melnik

Vita

Andrew currently works on Deep Reinforcement Learning approaches for robotics at CITEC, Bielefeld University, Germany.
He received his doctoral degree in Cognitive Science at Osnabrück University for his thesis on sensorimotor processing in the human brain and in cognitive architectures.

Website

https://youtu.be/XjsY8-P4WHM?t=48s