Timezone: »
We introduce a method for policy improvement that interpolates between the greedy approach of value-based reinforcement learning (RL) and the full planning approach typical of model-based RL. The new method builds on the concept of a geometric horizon model (GHM, also known as a \gamma-model), which models the discounted state-visitation distribution of a given policy. We show that we can evaluate any non-Markov policy that switches between a set of base Markov policies with fixed probability by a careful composition of the base policy GHMs, without any additional learning. We can then apply generalised policy improvement (GPI) to collections of such non-Markov policies to obtain a new Markov policy that will in general outperform its precursors. We provide a thorough theoretical analysis of this approach, develop applications to transfer and standard RL, and empirically demonstrate its effectiveness over standard GPI on a challenging deep RL continuous control task. We also provide an analysis of GHM training methods, proving a novel convergence result regarding previously proposed methods and showing how to train these models stably in deep RL settings.
Author Information
Shantanu Thakoor (DeepMind)
Mark Rowland (DeepMind)
Diana Borsa (DeepMind)
Will Dabney (DeepMind)
Remi Munos (DeepMind)
Andre Barreto (DeepMind)
Related Events (a corresponding poster, oral, or spotlight)
-
2022 Oral: Generalised Policy Improvement with Geometric Policy Composition »
Thu. Jul 21st 05:30 -- 05:50 PM Room Room 327 - 329
More from the Same Authors
-
2021 : Marginalized Operators for Off-Policy Reinforcement Learning »
Yunhao Tang · Mark Rowland · Remi Munos · Michal Valko -
2021 : Beyond Fine-Tuning: Transferring Behavior in Reinforcement Learning »
Víctor Campos · Pablo Sprechmann · Steven Hansen · Andre Barreto · Steven Kapturowski · Alex Vitvitskyi · Adrià Puigdomenech Badia · Charles Blundell -
2021 : Discovering Diverse Nearly Optimal Policies with Successor Features »
Tom Zahavy · Brendan O'Donoghue · Andre Barreto · Sebastian Flennerhag · Vlad Mnih · Satinder Singh -
2023 Poster: Bootstrapped Representations in Reinforcement Learning »
Charline Le Lan · Stephen Tu · Mark Rowland · Anna Harutyunyan · Rishabh Agarwal · Marc Bellemare · Will Dabney -
2023 Poster: The Statistical Benefits of Quantile Temporal-Difference Learning for Value Estimation »
Mark Rowland · Yunhao Tang · Clare Lyle · Remi Munos · Marc Bellemare · Will Dabney -
2023 Poster: Half-Hop: A graph upsampling approach for slowing down message passing »
Mehdi Azabou · Venkataramana Ganesh · Shantanu Thakoor · Chi-Heng Lin · Lakshmi Sathidevi · Ran Liu · Michal Valko · Petar Veličković · Eva Dyer -
2023 Poster: Quantile Credit Assignment »
Thomas Mesnard · Wenqi Chen · Alaa Saade · Yunhao Tang · Mark Rowland · Theophane Weber · Clare Lyle · Audrunas Gruslys · Michal Valko · Will Dabney · Georg Ostrovski · Eric Moulines · Remi Munos -
2023 Poster: Curiosity in Hindsight: Intrinsic Exploration in Stochastic Environments »
Daniel Jarrett · Corentin Tallec · Florent Altché · Thomas Mesnard · Remi Munos · Michal Valko -
2023 Poster: Adapting to game trees in zero-sum imperfect information games »
Côme Fiegel · Pierre Menard · Tadashi Kozuno · Remi Munos · Vianney Perchet · Michal Valko -
2023 Poster: Understanding Plasticity in Neural Networks »
Clare Lyle · Zeyu Zheng · Evgenii Nikishin · Bernardo Avila Pires · Razvan Pascanu · Will Dabney -
2023 Poster: Settling the Reward Hypothesis »
John Martin · Michael Bowling · David Abel · Will Dabney -
2023 Poster: Representations and Exploration for Deep Reinforcement Learning using Singular Value Decomposition »
Yash Chandak · Shantanu Thakoor · Zhaohan Guo · Yunhao Tang · Remi Munos · Will Dabney · Diana Borsa -
2023 Poster: Fast Rates for Maximum Entropy Exploration »
Daniil Tiapkin · Denis Belomestny · Daniele Calandriello · Eric Moulines · Remi Munos · Alexey Naumov · Pierre Perrault · Yunhao Tang · Michal Valko · Pierre Menard -
2023 Poster: DoMo-AC: Doubly Multi-step Off-policy Actor-Critic Algorithm »
Yunhao Tang · Tadashi Kozuno · Mark Rowland · Anna Harutyunyan · Remi Munos · Bernardo Avila Pires · Michal Valko -
2023 Poster: Understanding Self-Predictive Learning for Reinforcement Learning »
Yunhao Tang · Zhaohan Guo · Pierre Richemond · Bernardo Avila Pires · Yash Chandak · Remi Munos · Mark Rowland · Mohammad Gheshlaghi Azar · Charline Le Lan · Clare Lyle · Andras Gyorgy · Shantanu Thakoor · Will Dabney · Bilal Piot · Daniele Calandriello · Michal Valko -
2023 Poster: Towards a better understanding of representation dynamics under TD-learning »
Yunhao Tang · Remi Munos -
2023 Poster: VA-learning as a more efficient alternative to Q-learning »
Yunhao Tang · Remi Munos · Mark Rowland · Michal Valko -
2023 Poster: Regularization and Variance-Weighted Regression Achieves Minimax Optimality in Linear MDPs: Theory and Practice »
Toshinori Kitamura · Tadashi Kozuno · Yunhao Tang · Nino Vieillard · Michal Valko · Wenhao Yang · Jincheng Mei · Pierre Menard · Mohammad Gheshlaghi Azar · Remi Munos · Olivier Pietquin · Matthieu Geist · Csaba Szepesvari · Wataru Kumagai · Yutaka Matsuo -
2023 Oral: Quantile Credit Assignment »
Thomas Mesnard · Wenqi Chen · Alaa Saade · Yunhao Tang · Mark Rowland · Theophane Weber · Clare Lyle · Audrunas Gruslys · Michal Valko · Will Dabney · Georg Ostrovski · Eric Moulines · Remi Munos -
2023 Oral: Adapting to game trees in zero-sum imperfect information games »
Côme Fiegel · Pierre Menard · Tadashi Kozuno · Remi Munos · Vianney Perchet · Michal Valko -
2023 Oral: Understanding Plasticity in Neural Networks »
Clare Lyle · Zeyu Zheng · Evgenii Nikishin · Bernardo Avila Pires · Razvan Pascanu · Will Dabney -
2023 Oral: Settling the Reward Hypothesis »
John Martin · Michael Bowling · David Abel · Will Dabney -
2022 Workshop: Decision Awareness in Reinforcement Learning »
Evgenii Nikishin · Pierluca D'Oro · Doina Precup · Andre Barreto · Amir-massoud Farahmand · Pierre-Luc Bacon -
2022 Poster: Learning Dynamics and Generalization in Deep Reinforcement Learning »
Clare Lyle · Mark Rowland · Will Dabney · Marta Kwiatkowska · Yarin Gal -
2022 Spotlight: Learning Dynamics and Generalization in Deep Reinforcement Learning »
Clare Lyle · Mark Rowland · Will Dabney · Marta Kwiatkowska · Yarin Gal -
2022 Poster: Model-Value Inconsistency as a Signal for Epistemic Uncertainty »
Angelos Filos · Eszter Vértes · Zita Marinho · Gregory Farquhar · Diana Borsa · Abe Friesen · Feryal Behbahani · Tom Schaul · Andre Barreto · Simon Osindero -
2022 Spotlight: Model-Value Inconsistency as a Signal for Epistemic Uncertainty »
Angelos Filos · Eszter Vértes · Zita Marinho · Gregory Farquhar · Diana Borsa · Abe Friesen · Feryal Behbahani · Tom Schaul · Andre Barreto · Simon Osindero -
2021 Poster: Revisiting Peng's Q($\lambda$) for Modern Reinforcement Learning »
Tadashi Kozuno · Yunhao Tang · Mark Rowland · Remi Munos · Steven Kapturowski · Will Dabney · Michal Valko · David Abel -
2021 Poster: From Poincaré Recurrence to Convergence in Imperfect Information Games: Finding Equilibrium via Regularization »
Julien Perolat · Remi Munos · Jean-Baptiste Lespiau · Shayegan Omidshafiei · Mark Rowland · Pedro Ortega · Neil Burch · Thomas Anthony · David Balduzzi · Bart De Vylder · Georgios Piliouras · Marc Lanctot · Karl Tuyls -
2021 Poster: Taylor Expansion of Discount Factors »
Yunhao Tang · Mark Rowland · Remi Munos · Michal Valko -
2021 Spotlight: Taylor Expansion of Discount Factors »
Yunhao Tang · Mark Rowland · Remi Munos · Michal Valko -
2021 Spotlight: Revisiting Peng's Q($\lambda$) for Modern Reinforcement Learning »
Tadashi Kozuno · Yunhao Tang · Mark Rowland · Remi Munos · Steven Kapturowski · Will Dabney · Michal Valko · David Abel -
2021 Spotlight: From Poincaré Recurrence to Convergence in Imperfect Information Games: Finding Equilibrium via Regularization »
Julien Perolat · Remi Munos · Jean-Baptiste Lespiau · Shayegan Omidshafiei · Mark Rowland · Pedro Ortega · Neil Burch · Thomas Anthony · David Balduzzi · Bart De Vylder · Georgios Piliouras · Marc Lanctot · Karl Tuyls -
2021 Poster: Counterfactual Credit Assignment in Model-Free Reinforcement Learning »
Thomas Mesnard · Theophane Weber · Fabio Viola · Shantanu Thakoor · Alaa Saade · Anna Harutyunyan · Will Dabney · Thomas Stepleton · Nicolas Heess · Arthur Guez · Eric Moulines · Marcus Hutter · Lars Buesing · Remi Munos -
2021 Spotlight: Counterfactual Credit Assignment in Model-Free Reinforcement Learning »
Thomas Mesnard · Theophane Weber · Fabio Viola · Shantanu Thakoor · Alaa Saade · Anna Harutyunyan · Will Dabney · Thomas Stepleton · Nicolas Heess · Arthur Guez · Eric Moulines · Marcus Hutter · Lars Buesing · Remi Munos -
2020 Poster: Revisiting Fundamentals of Experience Replay »
William Fedus · Prajit Ramachandran · Rishabh Agarwal · Yoshua Bengio · Hugo Larochelle · Mark Rowland · Will Dabney -
2020 Poster: Fast computation of Nash Equilibria in Imperfect Information Games »
Remi Munos · Julien Perolat · Jean-Baptiste Lespiau · Mark Rowland · Bart De Vylder · Marc Lanctot · Finbarr Timbers · Daniel Hennes · Shayegan Omidshafiei · Audrunas Gruslys · Mohammad Gheshlaghi Azar · Edward Lockhart · Karl Tuyls -
2019 Poster: Statistics and Samples in Distributional Reinforcement Learning »
Mark Rowland · Robert Dadashi · Saurabh Kumar · Remi Munos · Marc Bellemare · Will Dabney -
2019 Oral: Statistics and Samples in Distributional Reinforcement Learning »
Mark Rowland · Robert Dadashi · Saurabh Kumar · Remi Munos · Marc Bellemare · Will Dabney -
2019 Poster: Composing Entropic Policies using Divergence Correction »
Jonathan Hunt · Andre Barreto · Timothy Lillicrap · Nicolas Heess -
2019 Oral: Composing Entropic Policies using Divergence Correction »
Jonathan Hunt · Andre Barreto · Timothy Lillicrap · Nicolas Heess -
2018 Poster: The Uncertainty Bellman Equation and Exploration »
Brendan O'Donoghue · Ian Osband · Remi Munos · Vlad Mnih -
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: Autoregressive Quantile Networks for Generative Modeling »
Georg Ostrovski · Will Dabney · Remi Munos -
2018 Oral: The Uncertainty Bellman Equation and Exploration »
Brendan O'Donoghue · Ian Osband · Remi Munos · Vlad Mnih -
2018 Oral: Autoregressive Quantile Networks for Generative Modeling »
Georg Ostrovski · Will Dabney · Remi Munos -
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 Poster: Implicit Quantile Networks for Distributional Reinforcement Learning »
Will Dabney · Georg Ostrovski · David Silver · Remi Munos -
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: Implicit Quantile Networks for Distributional Reinforcement Learning »
Will Dabney · Georg Ostrovski · David Silver · 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 Poster: Count-Based Exploration with Neural Density Models »
Georg Ostrovski · Marc Bellemare · Aäron van den Oord · Remi Munos -
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 Talk: Count-Based Exploration with Neural Density Models »
Georg Ostrovski · Marc Bellemare · Aäron van den Oord · Remi Munos -
2017 Poster: A Distributional Perspective on Reinforcement Learning »
Marc Bellemare · Will Dabney · Remi Munos -
2017 Poster: Automated Curriculum Learning for Neural Networks »
Alex Graves · Marc Bellemare · Jacob Menick · Remi Munos · Koray Kavukcuoglu -
2017 Poster: Minimax Regret Bounds for Reinforcement Learning »
Mohammad Gheshlaghi Azar · Ian Osband · Remi Munos -
2017 Talk: A Distributional Perspective on Reinforcement Learning »
Marc Bellemare · Will Dabney · Remi Munos -
2017 Talk: Automated Curriculum Learning for Neural Networks »
Alex Graves · Marc Bellemare · Jacob Menick · Remi Munos · Koray Kavukcuoglu -
2017 Talk: Minimax Regret Bounds for Reinforcement Learning »
Mohammad Gheshlaghi Azar · Ian Osband · Remi Munos