ECE 472 Robotic and Computer Vision Project 2: Deep Reinforcement Learning
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
ECE 472 Robotic and Computer Vision
Project 2: Deep Reinforcement Learning
1. Write a Jupyter notebook in Google Co-lab to implement the Pytorch tutorial cart-pole:
(https://pytorch.org/tutorials/intermediate/reinforcement q learning.html ). For each key line of code, please give a brief description of the RL algorithm
that is implemented (e.g. 2-3 sentences).
2. Give an overview of the DQN algorithm in written forms (1-3 paragraphs). Include deinitions of key terms and how they relate to the cart-pole prob- lem ( state, action, environment, reward. )
3. Choose performance metrics for your implementation and plot these as a function of the amount of training for your agent. Compare these plots for a random agent, and an agent trained for a small amount of time and a large amount of time. (Choose training times to show signiicant diferences in the plots).
4. Explore open-AI gym to ind three other problems that can be solved with RL. Compose a list of three other environments. Describe the state, action, environment and reward for each of these three environments.
Note: Displaying the cart-pole is somewhat tricky in google colab. There is a tutorial here that explains how it can be done:
https://www.anyscale.com/blog/an-introduction-to-reinforcement-learning- with-openai-gym-rllib-and-google
2023-11-18
Deep Reinforcement Learning