https://github.com/vdblm/human-machine-mdp
https://github.com/vdblm/human-machine-mdp
Last synced: 4 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/vdblm/human-machine-mdp
- Owner: vdblm
- Created: 2019-09-19T14:46:40.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2022-11-22T06:09:27.000Z (over 3 years ago)
- Last Synced: 2025-09-02T16:58:36.341Z (10 months ago)
- Language: Jupyter Notebook
- Size: 468 KB
- Stars: 2
- Watchers: 1
- Forks: 1
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Learning to Switch Between Machines and Humans
## Requirements
To install requirements:
```setup
pip install -r requirements.txt
```
## Code structure (Training and Evaluation)
- `agents/`
- `agents/hum_mac_agents.py` contains human/machine action policies.
- `agents/switching_agents.py` contains implementation of Algorithm 1, Algorithm 2, UCRL2, and the Greedy algorithm in the paper.
- `environments/` contains the code to produce all the environment types (episodic MDPs) used in the paper (i.e., Env-1, Env-2, and Env-3).
- `experiments/` contains all the known/unknown human experiments. For example,
- `SensorBasedSwitchingExperiment.run_unknown_human_exp` trains and evaluates Algorithm 2, UCRL2, and the Greedy algorithm in an episodic setting.
- `plot/` contains the code to plot the figures in the paper.
## Results
See the notebook `plots.ipynb` to reproduce the results in the paper:
```notebook
jupyter notebook plots.ipynb
```