Filters








280 Hits in 4.1 sec

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

Marlos C. Machado, Marc G. Bellemare, Erik Talvitie, Joel Veness, Matthew Hausknecht, Michael Bowling
2017 arXiv   pre-print
The Arcade Learning Environment (ALE) is an evaluation platform that poses the challenge of building AI agents with general competency across dozens of Atari 2600 games.  ...  We conclude this big picture look by revisiting challenges posed when the ALE was introduced, summarizing the state-of-the-art in various problems and highlighting problems that remain open.  ...  We would also like to thank the several contributors to the Arcade Learning Environment GitHub repository, specially Nicolas Carion for implementing the mode and difficult selection and Ben Goodrich for  ... 
arXiv:1709.06009v2 fatcat:f4zfxv73eja4pk4oirc4cwkugq

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

Marlos C. Machado, Marc G. Bellemare, Erik Talvitie, Joel Veness, Matthew Hausknecht, Michael Bowling
2018 The Journal of Artificial Intelligence Research  
The Arcade Learning Environment (ALE) is an evaluation platform that poses the challenge of building AI agents with general competency across dozens of Atari 2600 games.  ...  We conclude this big picture look by revisiting challenges posed when the ALE was introduced, summarizing the state-of-the-art in various problems and highlighting problems that remain open.  ...  Introduction The Arcade Learning Environment (ALE) is both a challenge problem and a platform for evaluating general competency in artificial intelligence (AI).  ... 
doi:10.1613/jair.5699 fatcat:a7634ndx3bdkfny7rafndkq4ai

Designing Online Environments for Expert/Novice Collaboration

José P. Zagal, Amy Bruckman
2010 Convergence. The International Journal of Research into New Media Technologies  
Students found that their participation was enjoyable and useful for learning. Also, there is evidence that they developed a deeper understanding of the medium of videogames.  ...  We conclude with thoughts on the importance of these kinds of authentic environments in traditional learning.  ...  Acknowledgements We would like to thank our anonymous reviewers for the insightful and constructive comments and feedback.  ... 
doi:10.1177/1354856510375141 fatcat:hiquplbumna4thfe3m7bsyy2ni

Authentic Learning Environments [chapter]

Jan Herrington, Thomas C. Reeves, Ron Oliver
2013 Handbook of Research on Educational Communications and Technology  
learning and real-life learning.  ...  One theory of learning which has the capacity to promote authentic learning is that of situated learning. vi Table of contents Abstract ii Declaration iv  ...  : agent, activity, and world.  ... 
doi:10.1007/978-1-4614-3185-5_32 fatcat:vvm2car6r5aulduwofly2u7o3u

A Study on Overfitting in Deep Reinforcement Learning [article]

Chiyuan Zhang and Oriol Vinyals and Remi Munos and Samy Bengio
2018 arXiv   pre-print
The observations call for more principled and careful evaluation protocols in RL.  ...  As deep RL techniques are being applied to critical problems such as healthcare and finance, it is important to understand the generalization behaviors of the trained agents.  ...  Acknowledgments The authors would like to thank Neil Rabinowitz, Eric Jang and David Silver for helpful discussions and comments.  ... 
arXiv:1804.06893v2 fatcat:74gd4s6f2zbmjd4r7a7n4z3nde

Leveraging Procedural Generation to Benchmark Reinforcement Learning [article]

Karl Cobbe, Christopher Hesse, Jacob Hilton, John Schulman
2020 arXiv   pre-print
We empirically demonstrate that diverse environment distributions are essential to adequately train and evaluate RL agents, thereby motivating the extensive use of procedural content generation.  ...  We believe that the community will benefit from increased access to high quality training environments, and we provide detailed experimental protocols for using this benchmark.  ...  Hausknecht, and M. Bowling. Revisiting the ar- cade learning environment: Evaluation protocols and open problems for general agents. Journal of Artificial Intelligence Research, 61:523-562, 2018.  ... 
arXiv:1912.01588v2 fatcat:um7sgficpbdnhg56wb5pju5upy

Towards Continual Reinforcement Learning: A Review and Perspectives [article]

Khimya Khetarpal, Matthew Riemer, Irina Rish, Doina Precup
2020 arXiv   pre-print
We go on to discuss evaluation of continual RL agents, providing an overview of benchmarks used in the literature and important metrics for understanding agent performance.  ...  Finally, we highlight open problems and challenges in bridging the gap between the current state of continual RL and findings in neuroscience.  ...  We would like to thank Takuya Ito and Martin Klissarov for providing valuable feedback.  ... 
arXiv:2012.13490v1 fatcat:vcleqjnpgrbkvg477d4prmzg2q

Prioritized Level Replay [article]

Minqi Jiang, Edward Grefenstette, Tim Rocktäschel
2021 arXiv   pre-print
Environments with procedurally generated content serve as important benchmarks for testing systematic generalization in deep reinforcement learning.  ...  We introduce Prioritized Level Replay (PLR), a general framework for selectively sampling the next training level by prioritizing those with higher estimated learning potential when revisited in the future  ...  Acknowledgements We thank Roberta Raileanu, Heinrich Küttler, and Jakob Foerster for useful discussions and feedback on this work, and our anonymous reviewers, for their recommendations on improving this  ... 
arXiv:2010.03934v4 fatcat:a4urejqeirf27adktdez5p6w6q

Improving Experience Replay through Modeling of Similar Transitions' Sets [article]

Daniel Eugênio Neves, João Pedro Oliveira Batisteli, Eduardo Felipe Lopes, Lucila Ishitani, Zenilton Kleber Gonçalves do Patrocínio Júnior
2021 arXiv   pre-print
We also present results for a DQN agent with the same experimental protocol on the same games set as the baseline.  ...  We report detailed results from five training trials of COMPER for just 100,000 frames and about 25,000 iterations with a small experiences memory on eight challenging games of Arcade Learning Environment  ...  Acknowledgments This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -Brasil (CAPES) -Finance Code 001.  ... 
arXiv:2111.06907v1 fatcat:flcjgnjnzfgqhhssl4735npafm

An Optimistic Perspective on Offline Reinforcement Learning [article]

Rishabh Agarwal, Dale Schuurmans, Mohammad Norouzi
2020 arXiv   pre-print
The DQN replay dataset can serve as an offline RL benchmark and is open-sourced.  ...  To enhance generalization in the offline setting, we present Random Ensemble Mixture (REM), a robust Q-learning algorithm that enforces optimal Bellman consistency on random convex combinations of multiple  ...  Acknowledgements We thank Pablo Samuel Castro for help in understanding and debugging issues with the Dopamine codebase and reviewing an early draft of the paper.  ... 
arXiv:1907.04543v4 fatcat:ocqec67o7zhvtlsz4sjlwvoa7e

Never Give Up: Learning Directed Exploration Strategies [article]

Adrià Puigdomènech Badia, Pablo Sprechmann, Alex Vitvitskyi, Daniel Guo, Bilal Piot, Steven Kapturowski, Olivier Tieleman, Martín Arjovsky, Alexander Pritzel, Andew Bolt, Charles Blundell
2020 arXiv   pre-print
revisit all states in its environment.  ...  and exploitation.  ...  ATARI RESULTS In this section, we evaluate the effectiveness of the NGU agent on the Arcade Learning Environment (ALE; (Bellemare et al., 2013) ).  ... 
arXiv:2002.06038v1 fatcat:jlhm6h4afnbyjhzdslyarwp2xy

Measuring Progress in Deep Reinforcement Learning Sample Efficiency [article]

Florian E. Dorner
2021 arXiv   pre-print
Current DRL benchmarks often allow for the cheap and easy generation of large amounts of samples such that perceived progress in DRL does not necessarily correspond to improved sample efficiency.  ...  Sampled environment transitions are a critical input to deep reinforcement learning (DRL) algorithms.  ...  Chris van Merwijk and the anonymous reviewers.  ... 
arXiv:2102.04881v1 fatcat:szmlx6yzl5ac3npmknmapghvay

A Survey and Critique of Multiagent Deep Reinforcement Learning [article]

Pablo Hernandez-Leal, Bilal Kartal, Matthew E. Taylor
2019 arXiv   pre-print
(ii) We provide general guidelines to new practitioners in the area: describing lessons learned from MDRL works, pointing to recent benchmarks, and outlining open avenues of research.  ...  Recent works have explored learning beyond single-agent scenarios and have considered multiagent learning (MAL) scenarios.  ...  for her visual designs for the figures in the article, to Frans Oliehoek, Sam Devlin, Marc Lanctot, Nolan Bard, Roberta Raileanu, Angeliki Lazaridou, and Yuhang Song for clarifications in their areas of  ... 
arXiv:1810.05587v2 fatcat:h4ei5zx2xfa7xocktlefjrvef4

Count-Based Exploration with the Successor Representation

Marlos C. Machado, Marc G. Bellemare, Michael Bowling
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper we introduce a simple approach for exploration in reinforcement learning (RL) that allows us to develop theoretically justified algorithms in the tabular case but that is also extendable  ...  Here we show that the norm of the SR, while it is being learned, can be used as a reward bonus to incentivize exploration.  ...  report for DQN MMC CTS and DQN MMC PIXELCNN , and Yuri Burda for providing us the data we used to compute the performance we report for RND in Atari 2600 games.  ... 
doi:10.1609/aaai.v34i04.5955 fatcat:g4wk2eacnfdfdnjnewmlisdjvq

AI Evaluation: past, present and future [article]

Jose Hernandez-Orallo
2016 arXiv   pre-print
We describe the limitations of the many evaluation settings and competitions in these three categories and propose several ideas for a more systematic and robust evaluation.  ...  We discuss several possibilities: the adaptation of cognitive tests used for humans and animals, the development of tests derived from algorithmic information theory or more general approaches under the  ...  processing required for the Arcade Learning Environment [12] or the General Video Game Competition [143]).  ... 
arXiv:1408.6908v3 fatcat:6g5h2nzaezey5a3qy3us7lnkvu
« Previous Showing results 1 — 15 out of 280 results