OFENet is a feature extractor network for low-dimensional data to improve performance of Reinforcement Learning. It can be combined with algorithms such as PPO, DDPG, TD3, and SAC.
This repository contains OFENet implementation, RL algorithms, and hyperparameters, which we used in our paper. We ran these codes on Ubuntu 18.04 & GeForce 1060.
See INSTALL.md.
See GETTING_STARTED.md).
Explain how others can test your code (eg run tests, demos, etc)
If you use the software, please cite the following (TR2020-083):
@inproceedings{ota2020can,
title={Can increasing input dimensionality improve deep reinforcement learning?},
author={Ota, Kei and Oiki, Tomoaki and Jha, Devesh and Mariyama, Toshisada and Nikovski, Daniel},
booktitle={International conference on machine learning},
pages={7424--7433},
year={2020},
organization={PMLR}
}
See CONTRIBUTING.md for our policy on contributions.
Released under AGPL-3.0-or-later
license, as found in the LICENSE.md file.
All files, except as noted below:
Copyright (C) 2020, 2023 Mitsubishi Electric Research Laboratories (MERL).
SPDX-License-Identifier: AGPL-3.0-or-later
tfrl
was adapted from https://github.com/deepmind/trfl (Apache-2.0
license as found
in LICENSES/Apache-2.0.txt).
teflon/policy/PPO.py
was adapted from https://github.com/keiohta/tf2rl (MIT
license as found
in LICENSES/MIT.txt).
util/gin_tf_external.py
and util/gin_utils.py
were adapted from https://github.com/google/gin-config (Apache-2.0
license as found in LICENSES/Apache-2.0.txt).
If you have problem running codes, please contact Kei Ota (ota.kei@ds.mitsubishielectric.co.jp) or Devesh K. Jha (jha@merl.com)