https://github.com/bstee615/gym-adv
Gym environments modified with adversarial agents
Science Score: 10.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
○codemeta.json file
-
○.zenodo.json file
-
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (5.3%) to scientific vocabulary
Last synced: 10 months ago
·
JSON representation
Repository
Gym environments modified with adversarial agents
Basic Info
- Host: GitHub
- Owner: bstee615
- Default Branch: master
- Homepage: https://arxiv.org/abs/1703.02702
- Size: 253 KB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Fork of lerrel/gym-adv
Created over 5 years ago
· Last pushed over 9 years ago
https://github.com/bstee615/gym-adv/blob/master/
> Under Development
# Gym environments with adversarial disturbance agents
This contains the adversarial environments used in our work on Robust Adversarial Reinforcement Learning ([RARL](https://arxiv.org/abs/1703.02702)). We heavily build on OpenAI Gym.
## Getting Started
The environments are based on the MuJoCo environments wrapped by OpenAI Gym's environments ([info](https://gym.openai.com/envs#mujoco)). For more information on OpenAI Gym environments refer to the [Gym webpage](https://gym.openai.com/).
Since these environments use the OpenAI pyhton bindings for the MuJoCo environments, you'll need to install `mujoco-py` following [this](https://github.com/openai/mujoco-py).
## Example
```python
import gym
E = gym.make('InvertedPendulumAdv-v1')
current_observation = E.reset()
# Set maximum adversary force
E.update_adversary(6)
# Get a sample action
u = E.sample_action()
# u.pro corresponds to protagonist action, while u.adv corresponds to the adversary's action
# Perform action
new_observation, reward, done, ~ = E.step(u)
```
## Contact
Lerrel Pinto -- lerrelpATcsDOTcmuDOTedu.
Owner
- Name: Benjamin Steenhoek
- Login: bstee615
- Kind: user
- Website: benjijang.com
- Repositories: 12
- Profile: https://github.com/bstee615
3rd year PhD student @ ISU. Interests and research: deep learning, program analysis