Noisy Networks for Exploration
Abstract
We introduce NoisyNet, a deep reinforcement learning agent with parametric noise added to its weights, and show that the induced stochasticity of the agent’s policy can be used to aid efficient exploration. The parameters of the noise are learned with gradient descent along with the remaining network weights. NoisyNet is straightforward to implement and adds little computational overhead. We find that replacing the conventional exploration heuristics for A3C, DQN and Dueling agents (entropy reward and epsilon-greedy respectively) with NoisyNet yields substantially higher scores for a wide range of Atari games, in some cases advancing the agent from sub to super-human performance.
Cite
Text
Fortunato et al. "Noisy Networks for Exploration." International Conference on Learning Representations, 2018.Markdown
[Fortunato et al. "Noisy Networks for Exploration." International Conference on Learning Representations, 2018.](https://mlanthology.org/iclr/2018/fortunato2018iclr-noisy/)BibTeX
@inproceedings{fortunato2018iclr-noisy,
title = {{Noisy Networks for Exploration}},
author = {Fortunato, Meire and Azar, Mohammad Gheshlaghi and Piot, Bilal and Menick, Jacob and Hessel, Matteo and Osband, Ian and Graves, Alex and Mnih, Volodymyr and Munos, Remi and Hassabis, Demis and Pietquin, Olivier and Blundell, Charles and Legg, Shane},
booktitle = {International Conference on Learning Representations},
year = {2018},
url = {https://mlanthology.org/iclr/2018/fortunato2018iclr-noisy/}
}