Timezone: »
Recent Reinforcement Learning (RL) algorithms making use of Kullback-Leibler (KL) regularization as a core component have shown outstanding performance. Yet, only little is understood theoretically about why KL regularization helps, so far. We study KL regularization within an approximate value iteration scheme and show that it implicitly averages q-values. Leveraging this insight, we provide a very strong performance bound, the very first to combine two desirable aspects: a linear dependency to the horizon (instead of quadratic) and an error propagation term involving an averaging effect of the estimation errors (instead of an accumulation effect). We also study the more general case of an additional entropy regularizer. The resulting abstract scheme encompasses many existing RL algorithms. Some of our assumptions do not hold with neural networks, so we complement this theoretical analysis with an extensive empirical study.
Author Information
Nino Vieillard (Google Brain)
Tadashi Kozuno (Okinawa Institute of Science and Technology)
Bruno Scherrer (INRIA)
Olivier Pietquin (Google Research Brain Team)
Remi Munos (DeepMind)
Matthieu Geist (Google Research, Brain Team)
Related Events (a corresponding poster, oral, or spotlight)
-
2020 Oral: Leverage the Average: an Analysis of KL Regularization in Reinforcement Learning »
Thu Dec 10th 02:30 -- 02:45 PM Room Orals & Spotlights: Reinforcement Learning
More from the Same Authors
-
2020 Poster: Munchausen Reinforcement Learning »
Nino Vieillard · Olivier Pietquin · Matthieu Geist -
2020 Poster: Bootstrap Your Own Latent - A New Approach to Self-Supervised Learning »
Jean-Bastien Grill · Florian Strub · Florent Altché · Corentin Tallec · Pierre Richemond · Elena Buchatskaya · Carl Doersch · Bernardo Avila Pires · Zhaohan Guo · Mohammad Gheshlaghi Azar · Bilal Piot · koray kavukcuoglu · Remi Munos · Michal Valko -
2020 Oral: Bootstrap Your Own Latent - A New Approach to Self-Supervised Learning »
Jean-Bastien Grill · Florian Strub · Florent Altché · Corentin Tallec · Pierre Richemond · Elena Buchatskaya · Carl Doersch · Bernardo Avila Pires · Zhaohan Guo · Mohammad Gheshlaghi Azar · Bilal Piot · koray kavukcuoglu · Remi Munos · Michal Valko -
2020 Poster: Fictitious Play for Mean Field Games: Continuous Time Analysis and Applications »
Sarah Perrin · Julien Perolat · Mathieu Lauriere · Matthieu Geist · Romuald Elie · Olivier Pietquin -
2019 Poster: Planning in entropy-regularized Markov decision processes and games »
Jean-Bastien Grill · Omar Darwiche Domingues · Pierre Menard · Remi Munos · Michal Valko -
2019 Poster: Budgeted Reinforcement Learning in Continuous State Space »
Nicolas Carrara · Edouard Leurent · Romain Laroche · Tanguy Urvoy · Odalric-Ambrym Maillard · Olivier Pietquin -
2019 Poster: Multiagent Evaluation under Incomplete Information »
Mark Rowland · Shayegan Omidshafiei · Karl Tuyls · Julien Perolat · Michal Valko · Georgios Piliouras · Remi Munos -
2019 Spotlight: Multiagent Evaluation under Incomplete Information »
Mark Rowland · Shayegan Omidshafiei · Karl Tuyls · Julien Perolat · Michal Valko · Georgios Piliouras · Remi Munos -
2019 Poster: Hindsight Credit Assignment »
Anna Harutyunyan · Will Dabney · Thomas Mesnard · Mohammad Gheshlaghi Azar · Bilal Piot · Nicolas Heess · Hado van Hasselt · Gregory Wayne · Satinder Singh · Doina Precup · Remi Munos -
2019 Spotlight: Hindsight Credit Assignment »
Anna Harutyunyan · Will Dabney · Thomas Mesnard · Mohammad Gheshlaghi Azar · Bilal Piot · Nicolas Heess · Hado van Hasselt · Gregory Wayne · Satinder Singh · Doina Precup · Remi Munos -
2018 Workshop: Visually grounded interaction and language »
Florian Strub · Harm de Vries · Erik Wijmans · Samyak Datta · Ethan Perez · Mateusz Malinowski · Stefan Lee · Peter Anderson · Aaron Courville · Jeremie MARY · Dhruv Batra · Devi Parikh · Olivier Pietquin · Chiori HORI · Tim Marks · Anoop Cherian -
2018 Poster: Optimistic optimization of a Brownian »
Jean-Bastien Grill · Michal Valko · Remi Munos -
2018 Poster: Multiple-Step Greedy Policies in Approximate and Online Reinforcement Learning »
Yonathan Efroni · Gal Dalal · Bruno Scherrer · Shie Mannor -
2018 Spotlight: Multiple-Step Greedy Policies in Approximate and Online Reinforcement Learning »
Yonathan Efroni · Gal Dalal · Bruno Scherrer · Shie Mannor -
2018 Poster: Actor-Critic Policy Optimization in Partially Observable Multiagent Environments »
Sriram Srinivasan · Marc Lanctot · Vinicius Zambaldi · Julien Perolat · Karl Tuyls · Remi Munos · Michael Bowling -
2017 Workshop: Visually grounded interaction and language »
Florian Strub · Harm de Vries · Abhishek Das · Satwik Kottur · Stefan Lee · Mateusz Malinowski · Olivier Pietquin · Devi Parikh · Dhruv Batra · Aaron Courville · Jeremie Mary -
2017 Poster: Is the Bellman residual a bad proxy? »
Matthieu Geist · Bilal Piot · Olivier Pietquin -
2017 Poster: Successor Features for Transfer in Reinforcement Learning »
Andre Barreto · Will Dabney · Remi Munos · Jonathan Hunt · Tom Schaul · David Silver · Hado van Hasselt -
2017 Spotlight: Successor Features for Transfer in Reinforcement Learning »
Andre Barreto · Will Dabney · Remi Munos · Jonathan Hunt · Tom Schaul · David Silver · Hado van Hasselt -
2017 Poster: Modulating early visual processing by language »
Harm de Vries · Florian Strub · Jeremie Mary · Hugo Larochelle · Olivier Pietquin · Aaron Courville -
2017 Spotlight: Modulating early visual processing by language »
Harm de Vries · Florian Strub · Jeremie Mary · Hugo Larochelle · Olivier Pietquin · Aaron Courville -
2016 Poster: Unifying Count-Based Exploration and Intrinsic Motivation »
Marc Bellemare · Sriram Srinivasan · Georg Ostrovski · Tom Schaul · David Saxton · Remi Munos -
2016 Poster: Memory-Efficient Backpropagation Through Time »
Audrunas Gruslys · Remi Munos · Ivo Danihelka · Marc Lanctot · Alex Graves -
2016 Poster: Blazing the trails before beating the path: Sample-efficient Monte-Carlo planning »
Jean-Bastien Grill · Michal Valko · Remi Munos -
2016 Oral: Blazing the trails before beating the path: Sample-efficient Monte-Carlo planning »
Jean-Bastien Grill · Michal Valko · Remi Munos -
2016 Poster: Safe and Efficient Off-Policy Reinforcement Learning »
Remi Munos · Tom Stepleton · Anna Harutyunyan · Marc Bellemare -
2015 Poster: Black-box optimization of noisy functions with unknown smoothness »
Jean-Bastien Grill · Michal Valko · Remi Munos · Remi Munos -
2014 Poster: Difference of Convex Functions Programming for Reinforcement Learning »
Bilal Piot · Matthieu Geist · Olivier Pietquin -
2014 Spotlight: Difference of Convex Functions Programming for Reinforcement Learning »
Bilal Piot · Matthieu Geist · Olivier Pietquin -
2012 Poster: Inverse Reinforcement Learning through Structured Classification »
Edouard Klein · Matthieu Geist · BILAL PIOT · Olivier Pietquin