PINs: Parameterized Indexed Value Function for Efficient Exploration in Reinforcement Learning

This is the reference implementation of the algorithm PINs in the paper. This repository is based on environments from bsuite. To install bsuite, please follow instructions at https://github.com/deepmind/bsuite

Instructions

To run PINs on Cartpole-Swingup with sparse rewards, do:
python bsuite_experiment.py

Output files will be written to result_path in bsuite_experiment.py by default. The file plot.py can be used to plot episodic reward after training.

Communication

If you have a problem running the code or spot a bug, please open an issue.
Please direct other correspondence to Tian Tan: tiantan@stanford.edu