Towards robust and domain agnostic reinforcement learning competitions
Not on the map. Why?
Abstract (arXiv)
Reinforcement learning competitions have formed the basis for standard research benchmarks, galvanized advances in the state-of-the-art, and shaped the direction of the field. Despite this, a majority of challenges suffer from the same fundamental problems: participant solutions to the posed challenge are usually domain-specific, biased to maximally exploit compute resources, and not guaranteed to be reproducible. In this paper, we present a new framework of competition design that promotes the development of algorithms that overcome these barriers. We propose four central mechanisms for achieving this end: submission retraining, domain randomization, desemantization through domain obfuscation, and the limitation of competition compute and environment-sample budget. To demonstrate the efficacy of this design, we proposed, organized, and ran the MineRL 2020 Competition on Sample-Efficient Reinforcement Learning. In this work, we describe the organizational outcomes of the competition and show that the resulting participant submissions are reproducible, non-specific to the competition environment, and sample/resource efficient, despite the difficult competition task.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 40 of its 64 references that are in the map. First posted on arXiv 7 Jun 2021.
Builds on
40 of 64 references.
References · 40
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
- 2020Conservative Q-Learning for Offline Reinforcement Learningcited by 1,777
- 2015Deep Reinforcement Learning with Double Q-learningcited by 1,704
- 2018Off-Policy Deep Reinforcement Learning without Explorationcited by 1,227
- 2019Dota 2 with Large Scale Deep Reinforcement Learningcited by 1,084
- 2018Experience Replay for Continual Learningcited by 1,074
- 2015Prioritized Experience Replaycited by 1,070
- 2018IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architecturescited by 866
- 2017Rainbow: Combining Improvements in Deep Reinforcement Learningcited by 774
- 2015Dueling Network Architectures for Deep Reinforcement Learningcited by 769
- 2020Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixelscited by 573
28 more
- 2018Quantifying Generalization in Reinforcement Learningcited by 345
- 2017Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agentscited by 267
- 2020Decoupling Representation Learning from Reinforcement Learningcited by 263
- 2019MineRL: A Large-Scale Dataset of Minecraft Demonstrationscited by 210
- 2019Fixup Initialization: Residual Learning Without Normalizationcited by 194
- 2018Sample-Efficient Reinforcement Learning with Stochastic Ensemble Value Expansioncited by 167
- 2016A Deep Hierarchical Approach to Lifelong Learning in Minecraftcited by 90
- 2018A Dissection of Overfitting and Generalization in Continuous Reinforcement Learningcited by 86
- 2017Hierarchical and Interpretable Skill Acquisition in Multi-task Reinforcement Learningcited by 77
- 2016Control of Memory, Active Perception, and Action in Minecraftcited by 61
- 2018Gotta Learn Fast: A New Benchmark for Generalization in RLcited by 59
- 2019Obstacle Tower: A Generalization Challenge in Vision, Control, and Planningcited by 54
- 2021The MineRL 2020 Competition on Sample Efficient Reinforcement Learning using Human Priorscited by 40
- 2018Investigating Human Priors for Playing Video Gamescited by 39
- 2018Reinforced Imitation: Sample Efficient Deep Reinforcement Learning for Map-less Navigation by Leveraging Prior Demonstrationscited by 35
- 2020Flatland-RL : Multi-Agent Reinforcement Learning on Trainscited by 32
- 2020Sample-Efficient Automated Deep Reinforcement Learningcited by 31
- 2021When Is Generalizable Reinforcement Learning Tractable?cited by 19
- 2020Retrospective Analysis of the 2019 MineRL Competition on Sample Efficient Reinforcement Learningcited by 18
- 2020Sample Efficient Reinforcement Learning through Learning from Demonstrations in Minecraftcited by 15
- 2019The Multi-Agent Reinforcement Learning in Malm"O (MARL"O) Competitioncited by 14
- 2020Scaling Imitation Learning in Minecraftcited by 12
- 2020Solving Physics Puzzles by Reasoning about Pathscited by 9
- 2019Skynet: A Top Deep RL Agent in the Inaugural Pommerman Team Competitioncited by 7
- 2019Modularization of End-to-End Learning: Case Study in Arcade Gamescited by 6
- 2020Guaranteeing Reproducibility in Deep Learning Competitionscited by 4
- 2019Accelerating Training in Pommerman with Imitation and Reinforcement Learningcited by 3
- 2019Winning the ICCV 2019 Learning to Drive Challengecited by 1
Built on by
Citing papers per year: 2021: 2.
2 papers.