Interactive Reward Shaping

The code for the paper ``Iterative Reward Shaping using Human Feedback for Correcting Reward Misspecification''.

Installation

conda create -n irs python=3.7  
conda activate irs  
pip install -r requirements.txt  

Running Experiments

python main.py --task gridworld  

python main.py --task highway  

python main.py --task inventory