# MinAtar: An Atari-Inspired Testbed for Thorough and Reproducible Reinforcement Learning Experiments

arXiv: https://arxiv.org/abs/1903.03176 · HTML: https://mlsanity.com/paper/1903.03176/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Mar 2019.

## Built on by (most cited first, 19 of 19)

- [Greedification Operators for Policy Optimization: Investigating Forward and Reverse KL Divergences](https://mlsanity.com/paper/2107.08285/) (2021)
- [Mitigating Plasticity Loss in Continual Reinforcement Learning by Reducing Churn](https://mlsanity.com/paper/2506.00592/) (2025)
- [Improving Deep Reinforcement Learning by Reducing the Chain Effect of Value and Policy Churn](https://mlsanity.com/paper/2409.04792/) (2024)
- [Action Candidate Driven Clipped Double Q-learning for Discrete and Continuous Action Tasks](https://mlsanity.com/paper/2203.11526/) (2022)
- [Memory-efficient Reinforcement Learning with Value-based Knowledge Consolidation](https://mlsanity.com/paper/2205.10868/) (2022)
- [Action Candidate Based Clipped Double Q-learning for Discrete and Continuous Action Tasks](https://mlsanity.com/paper/2105.00704/) (2021)
- [Handling Delay in Real-Time Reinforcement Learning](https://mlsanity.com/paper/2503.23478/) (2025)
- [Balanced Q-learning: Combining the Influence of Optimistic and Pessimistic Targets](https://mlsanity.com/paper/2111.02787/) (2021)
- [Grid-to-Graph: Flexible Spatial Relational Inductive Biases for Reinforcement Learning](https://mlsanity.com/paper/2102.04220/) (2021)
- [Automating Control of Overestimation Bias for Reinforcement Learning](https://mlsanity.com/paper/2110.13523/) (2021)
- [Heuresis: Search Strategies for Autonomous AI Research Agents Across Quality, Diversity and Novelty](https://mlsanity.com/paper/2606.25198/) (2026)
- [Smoothing Advantage Learning](https://mlsanity.com/paper/2203.10445/) (2022)
- [Revisiting Adam for Streaming Reinforcement Learning](https://mlsanity.com/paper/2605.06764/) (2026)
- [Greedy-Step Off-Policy Reinforcement Learning](https://mlsanity.com/paper/2102.11717/) (2021)
- [Investigating the Edge of Stability Phenomenon in Reinforcement Learning](https://mlsanity.com/paper/2307.04210/) (2023)
- [Identifiable Token Correspondence for World Models](https://mlsanity.com/paper/2605.16457/) (2026)
- [Gimitest: A Comprehensive Tool for Testing Reinforcement Learning Policies](https://mlsanity.com/paper/2607.07029/) (2026)
- [Self-discovering RL in the Era of Experience: Is Learning History an Asset or a Burden?](https://mlsanity.com/paper/2609.35897/) (2026)
- [Common-Mode Errors Limit Low-Timestep Deep Spiking Q-Networks](https://mlsanity.com/paper/2610.07808/) (2026)

## Abstract (arXiv)

The Arcade Learning Environment (ALE) is a popular platform for evaluating reinforcement learning agents. Much of the appeal comes from the fact that Atari games demonstrate aspects of competency we expect from an intelligent agent and are not biased toward any particular solution approach. The challenge of the ALE includes (1) the representation learning problem of extracting pertinent information from raw pixels, and (2) the behavioural learning problem of leveraging complex, delayed associations between actions and rewards. Often, the research questions we are interested in pertain more to the latter, but the representation learning problem adds significant computational expense. We introduce MinAtar, short for miniature Atari, a new set of environments that capture the general mechanics of specific Atari games while simplifying the representational complexity to focus more on the behavioural challenges. MinAtar consists of analogues of five Atari games: Seaquest, Breakout, Asterix, Freeway and Space Invaders. Each MinAtar environment provides the agent with a 10x10xn binary state representation. Each game plays out on a 10x10 grid with n channels corresponding to game-specific objects, such as ball, paddle and brick in the game Breakout. To investigate the behavioural challenges posed by MinAtar, we evaluated a smaller version of the DQN architecture as well as online actor-critic with eligibility traces. With the representation learning problem simplified, we can perform experiments with significantly less computational expense. In our experiments, we use the saved compute time to perform step-size parameter sweeps and more runs than is typical for the ALE. Experiments like this improve reproducibility, and allow us to draw more confident conclusions. We hope that MinAtar can allow researchers to thoroughly investigate behavioural challenges similar to those inherent in the ALE.