Intervention-RL is a reinforcement learning (RL) framework designed for developing and evaluating interventions in complex environments. It provides a set of tools and algorithms to train RL agents and study their behavior under different intervention strategies.
- Clone the repository:
git clone https://github.com/username/intervention-rl.git - Install the Package:
pip install -e . - Install the required dependencies:
conda env create -f environment.yml
- Navigate to the project directory:
cd intervention-rl - Run the main script:
python -m scripts.train algo.a2c.exp_type="none" - Choose between the following intervention strategies:
"none": No blocker"expert_hirl": Algorithmic blocker, uses robot actions in buffer"expert_hirl": Algorithmic blocker, uses human actions in buffer"hirl": Uses trained blocker, without reward bonus"ours": Uses trained blocker, with reward bonus
- Customize configuration variables:
python -m scripts.train algo.a2c.exp_type="ours" seed=42 algo.a2c.learning_rate=0.001 env.catastrophe_clearance=8 env.blocker_clearance=8