https://github.com/awjuliani/meta-rl

Implementation of Meta-RL A3C algorithm
https://github.com/awjuliani/meta-rl

reinforcement-learning tensorflow

Last synced: about 15 hours ago
JSON representation

Implementation of Meta-RL A3C algorithm

• Host: GitHub
• URL: https://github.com/awjuliani/meta-rl
• Owner: awjuliani
• License: mit
• Created: 2017-01-28T21:43:45.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-02-22T17:12:45.000Z (over 7 years ago)
• Last Synced: 2024-08-08T23:19:48.916Z (3 months ago)
• Topics: reinforcement-learning, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 814 KB
• Stars: 401
• Watchers: 27
• Forks: 109
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Meta-RL

Tensorflow implementation of Meta-RL A3C algorithm taken from [Learning to Reinforcement Learn](https://arxiv.org/abs/1611.05763). 

For more information, as well as explainations of each of the experiments, see my corresponding [Medium post](https://medium.com/p/b15b592a2ddf). A3C is built from previous implementation available [here](https://github.com/awjuliani/DeepRL-Agents).

Contains iPython notebooks for:

* **A3C-Meta-Bandit** - Set of bandit tasks described in paper. Including: Independent, Dependent, and Restless bandits.

* **A3C-Meta-Context** - Rainbow bandit task using randomized colors to indicate reward-giving arm in each episode. 

* **A3C-Meta-Grid** - Rainbow Gridworld task; a variation of gridworld in which goal colors are randomzied each episode and must be learned "on the fly."