| Revisiting the arcade learning environment: Evaluation protocols and open problems for general agents MC Machado, MG Bellemare, E Talvitie, J Veness, M Hausknecht, ... Journal of Artificial Intelligence Research 61, 523-562, 2018 | 262 | 2018 |
| State of the art control of atari games using shallow reinforcement learning Y Liang, MC Machado, E Talvitie, M Bowling arXiv preprint arXiv:1512.01563, 2015 | 98 | 2015 |
| An Experts Algorithm for Transfer Learning. E Talvitie, S Singh International Joint Conference on Artificial Intelligence, 1065-1070, 2007 | 56 | 2007 |
| Model Regularization for Stable Sample Rollouts E Talvitie Uncertainty in Artificial Intelligence, 2014 | 53 | 2014 |
| Self-Correcting Models for Model-Based Reinforcement Learning E Talvitie Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence …, 2017 | 47 | 2017 |
| Skip Context Tree Switching MG Bellemare, J Veness, E Talvitie International Conference on Machine Learning, 2014 | 38 | 2014 |
| The effect of planning shape on dyna-style planning in high-dimensional state spaces GZ Holland, EJ Talvitie, M Bowling arXiv preprint arXiv:1806.01825, 2018 | 26 | 2018 |
| Simple local models for complex dynamical systems E Talvitie, S Singh Advances in Neural Information Processing Systems, 1617-1624, 2009 | 20 | 2009 |
| Learning to Make Predictions In Partially Observable Environments Without a Generative Model. E Talvitie, S Singh Journal of Artificial Intelligence Research (JAIR) 42, 353-392, 2011 | 19 | 2011 |
| Agnostic system identification for monte carlo planning E Talvitie Proceedings of the AAAI Conference on Artificial Intelligence 29 (1), 2015 | 18 | 2015 |
| Improving exploration in UCT using local manifolds S Srinivasan, E Talvitie, M Bowling Twenty-Ninth AAAI Conference on Artificial Intelligence, 2015 | 10 | 2015 |
| Learning partially observable models using temporally abstract decision trees E Talvitie Advances in Neural Information Processing Systems, 818-826, 2012 | 7 | 2012 |
| Policy Tree: Adaptive Representation for Policy Gradient UD Gupta, E Talvitie, M Bowling Twenty-Ninth AAAI Conference on Artificial Intelligence, 2015 | 6 | 2015 |
| Maintaining Predictions over Time without a Model. E Talvitie, S Singh International Joint Conference on Artificial Intelligence, 1249-1254, 2009 | 6 | 2009 |
| Simple Partial Models for Complex Dynamical Systems E Talvitie, S Singh The University of Michigan, 2010 | 5 | 2010 |
| Building Incomplete but Accurate Models. E Talvitie, B Wolfe, S Singh International Symposium on Artificial Intelligence and Mathematics, 2008 | 5 | 2008 |
| Selective Dyna-style planning under limited model capacity Z Abbas, S Sokota, E Talvitie, M White International Conference on Machine Learning, 1-10, 2020 | 4 | 2020 |
| Learning the reward function for a misspecified model E Talvitie International Conference on Machine Learning, 4838-4847, 2018 | 4 | 2018 |
| Pairwise Relative Offset Features for Atari 2600 Games. E Talvitie, MH Bowling AAAI Workshop: Learning for General Competency in Video Games, 2015 | 4 | 2015 |
| Hallucinating value: A pitfall of dyna-style planning with imperfect environment models T Jafferjee, E Imani, E Talvitie, M White, M Bowling arXiv preprint arXiv:2006.04363, 2020 | 2 | 2020 |
Michael BowlingUniversity of Alberta / DeepMindAdresse e-mail validée de ualberta.ca
