Soft Actor-Critic for Discrete Action Settings

Petros Christodoulou
10/16/2019
cs.LGcs.AIstat.ML

Official code: p-christ/Deep-Reinforcement-Learning-Algorithms-with-PyTorch (5,935 stars, MIT, last commit 2024-07-25).

Abstract

Soft Actor-Critic is a state-of-the-art reinforcement learning algorithm for continuous action settings that is not applicable to discrete action settings. Many important settings involve discrete actions, however, and so here we derive an alternative version of the Soft Actor-Critic algorithm that is applicable to discrete action settings. We then show that, even without any hyperparameter tuning, it is competitive with the tuned model-free state-of-the-art on a selection of games from the Atari suite.

View on arXivView PDF

Code Implementations(1)

p-christ/Deep-Reinforcement-Learning-Algorithms-with-PyTorchOfficial100%
MIT

PyTorch implementations of deep reinforcement learning algorithms and environments

5,9351,203PythonSep 7, 20182 years agoMIT

Similar Papers

Soft Decomposed Policy-Critic: Bridging the Gap for Effective Continuous Control with Discrete RL

86% match2023

Revisiting the Gumbel-Softmax in MADDPG

85% match1 repo292023

Bench-MFG: A Benchmark Suite for Learning in Stationary Mean Field Games

85% match1 repo32026

Multi-Pass Q-Networks for Deep Reinforcement Learning with Parameterised Action Spaces

84% match1 repo2402019

Algorithmic Framework for Model-based Deep Reinforcement Learning with Theoretical Guarantees

84% match1 repo552018

Discrete Flow Matching for Offline-to-Online Reinforcement Learning

83% match2026

Cite this paper

@article{christodoulou2019soft,
  title  = {Soft Actor-Critic for Discrete Action Settings},
  author = {Petros Christodoulou},
  year   = {2019},
  eprint = {1910.07207},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1910.07207}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1910.07207.svg)](https://codeofpaper.com/paper/1910.07207)

Discussion