Timezone: »
We investigate the combination of actor-critic reinforcement learning algorithms with a uniform large-scale experience replay and propose solutions for two ensuing challenges: (a) efficient actor-critic learning with experience replay (b) the stability of off-policy learning where agents learn from other agents behaviour. To this end we analyze the bias-variance tradeoffs in V-trace, a form of importance sampling for actor-critic methods. Based on our analysis, we then argue for mixing experience sampled from replay with on-policy experience, and propose a new trust region scheme that scales effectively to data distributions where V-trace becomes unstable. We provide extensive empirical validation of the proposed solutions on DMLab-30 and further show the benefits of this setup in two training regimes for Atari: (1) a single agent is trained up until 200M environment frames per game (2) a population of agents is trained up until 200M environment frames each and may share experience. We demonstrate state-of-the-art data efficiency among model-free agents in both regimes.
Author Information
Simon Schmitt (DeepMind)
Matteo Hessel (Deep Mind)
Karen Simonyan (DeepMind)
More from the Same Authors
-
2021 Poster: Emphatic Algorithms for Deep Reinforcement Learning »
Ray Jiang · Tom Zahavy · Zhongwen Xu · Adam White · Matteo Hessel · Charles Blundell · Hado van Hasselt -
2021 Poster: High-Performance Large-Scale Image Recognition Without Normalization »
Andy Brock · Soham De · Samuel Smith · Karen Simonyan -
2021 Spotlight: High-Performance Large-Scale Image Recognition Without Normalization »
Andy Brock · Soham De · Samuel Smith · Karen Simonyan -
2021 Spotlight: Emphatic Algorithms for Deep Reinforcement Learning »
Ray Jiang · Tom Zahavy · Zhongwen Xu · Adam White · Matteo Hessel · Charles Blundell · Hado van Hasselt -
2021 Poster: Learning and Planning in Complex Action Spaces »
Thomas Hubert · Julian Schrittwieser · Ioannis Antonoglou · Mohammadamin Barekatain · Simon Schmitt · David Silver -
2021 Poster: Muesli: Combining Improvements in Policy Optimization »
Matteo Hessel · Ivo Danihelka · Fabio Viola · Arthur Guez · Simon Schmitt · Laurent Sifre · Theophane Weber · David Silver · Hado van Hasselt -
2021 Spotlight: Learning and Planning in Complex Action Spaces »
Thomas Hubert · Julian Schrittwieser · Ioannis Antonoglou · Mohammadamin Barekatain · Simon Schmitt · David Silver -
2021 Spotlight: Muesli: Combining Improvements in Policy Optimization »
Matteo Hessel · Ivo Danihelka · Fabio Viola · Arthur Guez · Simon Schmitt · Laurent Sifre · Theophane Weber · David Silver · Hado van Hasselt -
2020 Poster: What Can Learned Intrinsic Rewards Capture? »
Zeyu Zheng · Junhyuk Oh · Matteo Hessel · Zhongwen Xu · Manuel Kroiss · Hado van Hasselt · David Silver · Satinder Singh -
2018 Poster: IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures »
Lasse Espeholt · Hubert Soyer · Remi Munos · Karen Simonyan · Vlad Mnih · Tom Ward · Yotam Doron · Vlad Firoiu · Tim Harley · Iain Dunning · Shane Legg · Koray Kavukcuoglu -
2018 Poster: Parallel WaveNet: Fast High-Fidelity Speech Synthesis »
Aäron van den Oord · Yazhe Li · Igor Babuschkin · Karen Simonyan · Oriol Vinyals · Koray Kavukcuoglu · George van den Driessche · Edward Lockhart · Luis C Cobo · Florian Stimberg · Norman Casagrande · Dominik Grewe · Seb Noury · Sander Dieleman · Erich Elsen · Nal Kalchbrenner · Heiga Zen · Alex Graves · Helen King · Tom Walters · Dan Belov · Demis Hassabis -
2018 Poster: Efficient Neural Audio Synthesis »
Nal Kalchbrenner · Erich Elsen · Karen Simonyan · Seb Noury · Norman Casagrande · Edward Lockhart · Florian Stimberg · Aäron van den Oord · Sander Dieleman · Koray Kavukcuoglu -
2018 Oral: Parallel WaveNet: Fast High-Fidelity Speech Synthesis »
Aäron van den Oord · Yazhe Li · Igor Babuschkin · Karen Simonyan · Oriol Vinyals · Koray Kavukcuoglu · George van den Driessche · Edward Lockhart · Luis C Cobo · Florian Stimberg · Norman Casagrande · Dominik Grewe · Seb Noury · Sander Dieleman · Erich Elsen · Nal Kalchbrenner · Heiga Zen · Alex Graves · Helen King · Tom Walters · Dan Belov · Demis Hassabis -
2018 Oral: Efficient Neural Audio Synthesis »
Nal Kalchbrenner · Erich Elsen · Karen Simonyan · Seb Noury · Norman Casagrande · Edward Lockhart · Florian Stimberg · Aäron van den Oord · Sander Dieleman · Koray Kavukcuoglu -
2018 Oral: IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures »
Lasse Espeholt · Hubert Soyer · Remi Munos · Karen Simonyan · Vlad Mnih · Tom Ward · Yotam Doron · Vlad Firoiu · Tim Harley · Iain Dunning · Shane Legg · Koray Kavukcuoglu -
2018 Poster: Transfer in Deep Reinforcement Learning Using Successor Features and Generalised Policy Improvement »
Andre Barreto · Diana Borsa · John Quan · Tom Schaul · David Silver · Matteo Hessel · Daniel J. Mankowitz · Augustin Zidek · Remi Munos -
2018 Poster: Learning to search with MCTSnets »
Arthur Guez · Theophane Weber · Ioannis Antonoglou · Karen Simonyan · Oriol Vinyals · Daan Wierstra · Remi Munos · David Silver -
2018 Oral: Transfer in Deep Reinforcement Learning Using Successor Features and Generalised Policy Improvement »
Andre Barreto · Diana Borsa · John Quan · Tom Schaul · David Silver · Matteo Hessel · Daniel J. Mankowitz · Augustin Zidek · Remi Munos -
2018 Oral: Learning to search with MCTSnets »
Arthur Guez · Theophane Weber · Ioannis Antonoglou · Karen Simonyan · Oriol Vinyals · Daan Wierstra · Remi Munos · David Silver -
2017 Poster: The Predictron: End-To-End Learning and Planning »
David Silver · Hado van Hasselt · Matteo Hessel · Tom Schaul · Arthur Guez · Tim Harley · Gabriel Dulac-Arnold · David Reichert · Neil Rabinowitz · Andre Barreto · Thomas Degris -
2017 Talk: The Predictron: End-To-End Learning and Planning »
David Silver · Hado van Hasselt · Matteo Hessel · Tom Schaul · Arthur Guez · Tim Harley · Gabriel Dulac-Arnold · David Reichert · Neil Rabinowitz · Andre Barreto · Thomas Degris -
2017 Poster: Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders »
Cinjon Resnick · Adam Roberts · Jesse Engel · Douglas Eck · Sander Dieleman · Karen Simonyan · Mohammad Norouzi -
2017 Talk: Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders »
Cinjon Resnick · Adam Roberts · Jesse Engel · Douglas Eck · Sander Dieleman · Karen Simonyan · Mohammad Norouzi -
2017 Poster: Video Pixel Networks »
Nal Kalchbrenner · Karen Simonyan · Aäron van den Oord · Ivo Danihelka · Oriol Vinyals · Alex Graves · Koray Kavukcuoglu -
2017 Talk: Video Pixel Networks »
Nal Kalchbrenner · Karen Simonyan · Aäron van den Oord · Ivo Danihelka · Oriol Vinyals · Alex Graves · Koray Kavukcuoglu