Towards robust and domain agnostic reinforcement learning competitions

TL;DR

This paper introduces a new competition framework for reinforcement learning that emphasizes reproducibility, domain-agnostic solutions, and resource efficiency, demonstrated through the MineRL 2020 Competition.

Contribution

It proposes four mechanisms—retraining, domain randomization, obfuscation, and resource limits—to improve RL competition design and showcases their effectiveness in a real-world challenge.

Findings

Submissions became more reproducible and domain-agnostic.

Participants developed sample-efficient algorithms.

The competition successfully promoted robust RL solutions.

Abstract

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.