Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Michael Bowling

Michael Bowling

Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games


Feb 13, 2021
Dustin Morrill, Ryan D'Orazio, Marc Lanctot, James R. Wright, Michael Bowling, Amy Greenwald

Figure 1 for Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games
Figure 2 for Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games
Figure 3 for Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games
Figure 4 for Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games
* 32 pages, 6 figures 
  Access Paper or Ask Questions

Solving Common-Payoff Games with Approximate Policy Iteration


Jan 11, 2021
Samuel Sokota, Edward Lockhart, Finbarr Timbers, Elnaz Davoodi, Ryan D'Orazio, Neil Burch, Martin Schmid, Michael Bowling, Marc Lanctot

Figure 1 for Solving Common-Payoff Games with Approximate Policy Iteration
Figure 2 for Solving Common-Payoff Games with Approximate Policy Iteration
Figure 3 for Solving Common-Payoff Games with Approximate Policy Iteration
Figure 4 for Solving Common-Payoff Games with Approximate Policy Iteration
* AAAI 2021 
  Access Paper or Ask Questions

Hindsight and Sequential Rationality of Correlated Play


Dec 17, 2020
Dustin Morrill, Ryan D'Orazio, Reca Sarfati, Marc Lanctot, James R. Wright, Amy Greenwald, Michael Bowling

Figure 1 for Hindsight and Sequential Rationality of Correlated Play
Figure 2 for Hindsight and Sequential Rationality of Correlated Play
Figure 3 for Hindsight and Sequential Rationality of Correlated Play
Figure 4 for Hindsight and Sequential Rationality of Correlated Play
* Technical report for a paper to appear at AAAI 2021. 26 pages, 15 figures 
  Access Paper or Ask Questions

Useful Policy Invariant Shaping from Arbitrary Advice


Nov 02, 2020
Paniz Behboudian, Yash Satsangi, Matthew E. Taylor, Anna Harutyunyan, Michael Bowling

Figure 1 for Useful Policy Invariant Shaping from Arbitrary Advice
Figure 2 for Useful Policy Invariant Shaping from Arbitrary Advice
Figure 3 for Useful Policy Invariant Shaping from Arbitrary Advice
Figure 4 for Useful Policy Invariant Shaping from Arbitrary Advice
* 9 pages, 6 figures, Adaptive and Learning Agents (ALA) 2020 Workshop 
  Access Paper or Ask Questions

The Advantage Regret-Matching Actor-Critic


Aug 27, 2020
Audrūnas Gruslys, Marc Lanctot, Rémi Munos, Finbarr Timbers, Martin Schmid, Julien Perolat, Dustin Morrill, Vinicius Zambaldi, Jean-Baptiste Lespiau, John Schultz, Mohammad Gheshlaghi Azar, Michael Bowling, Karl Tuyls

Figure 1 for The Advantage Regret-Matching Actor-Critic
Figure 2 for The Advantage Regret-Matching Actor-Critic
Figure 3 for The Advantage Regret-Matching Actor-Critic
Figure 4 for The Advantage Regret-Matching Actor-Critic

  Access Paper or Ask Questions

Marginal Utility for Planning in Continuous or Large Discrete Action Spaces


Jun 17, 2020
Zaheen Farraz Ahmad, Levi H. S. Lelis, Michael Bowling

Figure 1 for Marginal Utility for Planning in Continuous or Large Discrete Action Spaces
Figure 2 for Marginal Utility for Planning in Continuous or Large Discrete Action Spaces
Figure 3 for Marginal Utility for Planning in Continuous or Large Discrete Action Spaces
Figure 4 for Marginal Utility for Planning in Continuous or Large Discrete Action Spaces

  Access Paper or Ask Questions

Sample-Efficient Model-based Actor-Critic for an Interactive Dialogue Task


Apr 28, 2020
Katya Kudashkina, Valliappa Chockalingam, Graham W. Taylor, Michael Bowling

Figure 1 for Sample-Efficient Model-based Actor-Critic for an Interactive Dialogue Task
Figure 2 for Sample-Efficient Model-based Actor-Critic for an Interactive Dialogue Task
Figure 3 for Sample-Efficient Model-based Actor-Critic for an Interactive Dialogue Task
Figure 4 for Sample-Efficient Model-based Actor-Critic for an Interactive Dialogue Task

  Access Paper or Ask Questions

Approximate exploitability: Learning a best response in large games


Apr 20, 2020
Finbarr Timbers, Edward Lockhart, Martin Schmid, Marc Lanctot, Michael Bowling


  Access Paper or Ask Questions

Alternative Function Approximation Parameterizations for Solving Games: An Analysis of $f$-Regression Counterfactual Regret Minimization


Dec 06, 2019
Ryan D'Orazio, Dustin Morrill, James R. Wright, Michael Bowling

Figure 1 for Alternative Function Approximation Parameterizations for Solving Games: An Analysis of $f$-Regression Counterfactual Regret Minimization
Figure 2 for Alternative Function Approximation Parameterizations for Solving Games: An Analysis of $f$-Regression Counterfactual Regret Minimization
Figure 3 for Alternative Function Approximation Parameterizations for Solving Games: An Analysis of $f$-Regression Counterfactual Regret Minimization
* 11 pages, includes appendix. arXiv admin note: text overlap with arXiv:1910.01706 
  Access Paper or Ask Questions

Low-Variance and Zero-Variance Baselines for Extensive-Form Games


Jul 22, 2019
Trevor Davis, Martin Schmid, Michael Bowling

Figure 1 for Low-Variance and Zero-Variance Baselines for Extensive-Form Games
Figure 2 for Low-Variance and Zero-Variance Baselines for Extensive-Form Games
Figure 3 for Low-Variance and Zero-Variance Baselines for Extensive-Form Games
* Under review for NeurIPS 2019 
  Access Paper or Ask Questions

Rethinking Formal Models of Partially Observable Multiagent Decision Making


Jun 26, 2019
Vojtěch Kovařík, Martin Schmid, Neil Burch, Michael Bowling, Viliam Lisý

Figure 1 for Rethinking Formal Models of Partially Observable Multiagent Decision Making
Figure 2 for Rethinking Formal Models of Partially Observable Multiagent Decision Making
Figure 3 for Rethinking Formal Models of Partially Observable Multiagent Decision Making
Figure 4 for Rethinking Formal Models of Partially Observable Multiagent Decision Making

  Access Paper or Ask Questions

Ease-of-Teaching and Language Structure from Emergent Communication


Jun 06, 2019
Fushan Li, Michael Bowling

Figure 1 for Ease-of-Teaching and Language Structure from Emergent Communication
Figure 2 for Ease-of-Teaching and Language Structure from Emergent Communication
Figure 3 for Ease-of-Teaching and Language Structure from Emergent Communication
Figure 4 for Ease-of-Teaching and Language Structure from Emergent Communication

  Access Paper or Ask Questions

The Hanabi Challenge: A New Frontier for AI Research


Feb 01, 2019
Nolan Bard, Jakob N. Foerster, Sarath Chandar, Neil Burch, Marc Lanctot, H. Francis Song, Emilio Parisotto, Vincent Dumoulin, Subhodeep Moitra, Edward Hughes, Iain Dunning, Shibl Mourad, Hugo Larochelle, Marc G. Bellemare, Michael Bowling

Figure 1 for The Hanabi Challenge: A New Frontier for AI Research
Figure 2 for The Hanabi Challenge: A New Frontier for AI Research
Figure 3 for The Hanabi Challenge: A New Frontier for AI Research
Figure 4 for The Hanabi Challenge: A New Frontier for AI Research
* 37 pages, 5 figures, submitted to Artificial Intelligence 
  Access Paper or Ask Questions

Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning


Nov 04, 2018
Jakob N. Foerster, Francis Song, Edward Hughes, Neil Burch, Iain Dunning, Shimon Whiteson, Matthew Botvinick, Michael Bowling

Figure 1 for Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning
Figure 2 for Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning
Figure 3 for Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning
Figure 4 for Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning

  Access Paper or Ask Questions

Actor-Critic Policy Optimization in Partially Observable Multiagent Environments


Oct 21, 2018
Sriram Srinivasan, Marc Lanctot, Vinicius Zambaldi, Julien Perolat, Karl Tuyls, Remi Munos, Michael Bowling

Figure 1 for Actor-Critic Policy Optimization in Partially Observable Multiagent Environments
Figure 2 for Actor-Critic Policy Optimization in Partially Observable Multiagent Environments
Figure 3 for Actor-Critic Policy Optimization in Partially Observable Multiagent Environments
Figure 4 for Actor-Critic Policy Optimization in Partially Observable Multiagent Environments
* NIPS 2018 
  Access Paper or Ask Questions

Generalization and Regularization in DQN


Sep 29, 2018
Jesse Farebrother, Marlos C. Machado, Michael Bowling

Figure 1 for Generalization and Regularization in DQN
Figure 2 for Generalization and Regularization in DQN
Figure 3 for Generalization and Regularization in DQN
Figure 4 for Generalization and Regularization in DQN

  Access Paper or Ask Questions

Solving Large Extensive-Form Games with Strategy Constraints


Sep 20, 2018
Trevor Davis, Kevin Waugh, Michael Bowling

Figure 1 for Solving Large Extensive-Form Games with Strategy Constraints
Figure 2 for Solving Large Extensive-Form Games with Strategy Constraints
Figure 3 for Solving Large Extensive-Form Games with Strategy Constraints
* Submitted to AAAI 2019 
  Access Paper or Ask Questions

Variance Reduction in Monte Carlo Counterfactual Regret Minimization (VR-MCCFR) for Extensive Form Games using Baselines


Sep 09, 2018
Martin Schmid, Neil Burch, Marc Lanctot, Matej Moravcik, Rudolf Kadlec, Michael Bowling

Figure 1 for Variance Reduction in Monte Carlo Counterfactual Regret Minimization (VR-MCCFR) for Extensive Form Games using Baselines
Figure 2 for Variance Reduction in Monte Carlo Counterfactual Regret Minimization (VR-MCCFR) for Extensive Form Games using Baselines
Figure 3 for Variance Reduction in Monte Carlo Counterfactual Regret Minimization (VR-MCCFR) for Extensive Form Games using Baselines
Figure 4 for Variance Reduction in Monte Carlo Counterfactual Regret Minimization (VR-MCCFR) for Extensive Form Games using Baselines

  Access Paper or Ask Questions

Count-Based Exploration with the Successor Representation


Aug 14, 2018
Marlos C. Machado, Marc G. Bellemare, Michael Bowling

Figure 1 for Count-Based Exploration with the Successor Representation
Figure 2 for Count-Based Exploration with the Successor Representation
Figure 3 for Count-Based Exploration with the Successor Representation
Figure 4 for Count-Based Exploration with the Successor Representation

  Access Paper or Ask Questions

The Effect of Planning Shape on Dyna-style Planning in High-dimensional State Spaces


Jun 08, 2018
G. Zacharias Holland, Erik Talvitie, Michael Bowling

Figure 1 for The Effect of Planning Shape on Dyna-style Planning in High-dimensional State Spaces
Figure 2 for The Effect of Planning Shape on Dyna-style Planning in High-dimensional State Spaces

  Access Paper or Ask Questions

Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents


Dec 01, 2017
Marlos C. Machado, Marc G. Bellemare, Erik Talvitie, Joel Veness, Matthew Hausknecht, Michael Bowling

Figure 1 for Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents
Figure 2 for Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents
Figure 3 for Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents
Figure 4 for Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents

  Access Paper or Ask Questions

A Laplacian Framework for Option Discovery in Reinforcement Learning


Jun 16, 2017
Marlos C. Machado, Marc G. Bellemare, Michael Bowling

Figure 1 for A Laplacian Framework for Option Discovery in Reinforcement Learning
Figure 2 for A Laplacian Framework for Option Discovery in Reinforcement Learning
Figure 3 for A Laplacian Framework for Option Discovery in Reinforcement Learning
Figure 4 for A Laplacian Framework for Option Discovery in Reinforcement Learning
* Appearing in the Proceedings of the 34th International Conference on Machine Learning (ICML) 
  Access Paper or Ask Questions

DeepStack: Expert-Level Artificial Intelligence in No-Limit Poker


Mar 03, 2017
Matej Moravčík, Martin Schmid, Neil Burch, Viliam Lisý, Dustin Morrill, Nolan Bard, Trevor Davis, Kevin Waugh, Michael Johanson, Michael Bowling


  Access Paper or Ask Questions

AIVAT: A New Variance Reduction Technique for Agent Evaluation in Imperfect Information Games


Jan 19, 2017
Neil Burch, Martin Schmid, Matej Moravčík, Michael Bowling

Figure 1 for AIVAT: A New Variance Reduction Technique for Agent Evaluation in Imperfect Information Games
Figure 2 for AIVAT: A New Variance Reduction Technique for Agent Evaluation in Imperfect Information Games
Figure 3 for AIVAT: A New Variance Reduction Technique for Agent Evaluation in Imperfect Information Games
Figure 4 for AIVAT: A New Variance Reduction Technique for Agent Evaluation in Imperfect Information Games
* To appear at AAAI-17 Workshop on Computer Poker and Imperfect Information Games 
  Access Paper or Ask Questions

Learning Purposeful Behaviour in the Absence of Rewards


May 25, 2016
Marlos C. Machado, Michael Bowling

Figure 1 for Learning Purposeful Behaviour in the Absence of Rewards
Figure 2 for Learning Purposeful Behaviour in the Absence of Rewards
* Extended version of the paper presented at the workshop entitled Abstraction in Reinforcement Learning, at the 33rd International Conference on Machine Learning, New York, NY, USA, 2016 
  Access Paper or Ask Questions

State of the Art Control of Atari Games Using Shallow Reinforcement Learning


Apr 21, 2016
Yitao Liang, Marlos C. Machado, Erik Talvitie, Michael Bowling

Figure 1 for State of the Art Control of Atari Games Using Shallow Reinforcement Learning
Figure 2 for State of the Art Control of Atari Games Using Shallow Reinforcement Learning
Figure 3 for State of the Art Control of Atari Games Using Shallow Reinforcement Learning
Figure 4 for State of the Art Control of Atari Games Using Shallow Reinforcement Learning
* A shorter version of this paper appears in the Proceedings of the 15th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2016) 
  Access Paper or Ask Questions

Solving Games with Functional Regret Estimation


Dec 31, 2014
Kevin Waugh, Dustin Morrill, J. Andrew Bagnell, Michael Bowling

Figure 1 for Solving Games with Functional Regret Estimation
Figure 2 for Solving Games with Functional Regret Estimation
Figure 3 for Solving Games with Functional Regret Estimation
* AAAI Conference on Artificial Intelligence 2015 
  Access Paper or Ask Questions

Domain-Independent Optimistic Initialization for Reinforcement Learning


Oct 16, 2014
Marlos C. Machado, Sriram Srinivasan, Michael Bowling

Figure 1 for Domain-Independent Optimistic Initialization for Reinforcement Learning

  Access Paper or Ask Questions