English

The Dormant Neuron Phenomenon in Deep Reinforcement Learning

Machine Learning 2023-06-14 v2

Abstract

In this work we identify the dormant neuron phenomenon in deep reinforcement learning, where an agent's network suffers from an increasing number of inactive neurons, thereby affecting network expressivity. We demonstrate the presence of this phenomenon across a variety of algorithms and environments, and highlight its effect on learning. To address this issue, we propose a simple and effective method (ReDo) that Recycles Dormant neurons throughout training. Our experiments demonstrate that ReDo maintains the expressive power of networks by reducing the number of dormant neurons and results in improved performance.

Keywords

Cite

@article{arxiv.2302.12902,
  title  = {The Dormant Neuron Phenomenon in Deep Reinforcement Learning},
  author = {Ghada Sokar and Rishabh Agarwal and Pablo Samuel Castro and Utku Evci},
  journal= {arXiv preprint arXiv:2302.12902},
  year   = {2023}
}

Comments

Oral at ICML 2023

R2 v1 2026-06-28T08:49:11.947Z