Craftium: Bridging Flexibility and Efficiency for Rich 3D Single- and Multi-Agent Environments
Abstract: Advances in large models, reinforcement learning, and open-endedness have accelerated progress toward autonomous agents that can learn and interact in the real world. To achieve this, flexible tools are needed to create rich, yet computationally efficient, environments. While scalable 2D environments fail to address key real-world challenges like 3D navigation and spatial reasoning, more complex 3D environments are computationally expensive and lack features like customizability and multi-agent support. This paper introduces Craftium, a highly customizable and easy-to-use platform for building rich 3D single- and multi-agent environments. We showcase environments of different complexity and nature: from single- and multi-agent tasks to vast worlds with many creatures and biomes, and customizable procedural task generators. Benchmarking shows that Craftium significantly reduces the computational cost of alternatives of similar richness, achieving +2K steps per second more than Minecraft-based frameworks.
- Scaling instructable agents across many simulated worlds. arXiv preprint arXiv:2404.10179, 2024.
- Griddly: A platform for ai research in games. arXiv preprint arXiv:2011.06363, 2020.
- Human-timescale adaptation in an open-ended task space. In International Conference on Machine Learning (ICML) of 2023, volume 202, pages 1887–1935. PMLR, 2023.
- Deepmind Lab. arXiv preprint arXiv:1612.03801, 2016.
- The Arcade Learning Environment: An evaluation platform for general agents. Journal of Artificial Intelligence Research, 47:253–279, 2013.
- Proximal and Sparse Resolution of Constrained Dynamic Equations. In Robotics: Science and Systems of 2021, 2021.
- Minigrid & miniworld: Modular & customizable reinforcement learning environments for goal-oriented tasks. arXiv preprint arXiv:2306.13831, 2023.
- Leveraging procedural generation to benchmark reinforcement learning. arXiv preprint arXiv:1912.01588, 2020.
- CARLA: An open urban driving simulator. In the 1st Annual Conference on Robot Learning (CoRL), pages 1–16, 2017.
- MineDojo: Building open-ended embodied agents with internet-scale knowledge. Advances in Neural Information Processing Systems (NeurIPS) of 2022, 35:18343–18362, 2022.
- panda-gym: Open-Source Goal-Conditioned Environments for Robotic Learning. 4th Robot Learning Workshop: Self-Supervised and Lifelong Learning at NeurIPS, 2021.
- MineRL: A large-scale dataset of minecraft demonstrations. arXiv preprint arXiv:1907.13440, 2019.
- CleanRL: High-quality single-file implementations of deep reinforcement learning algorithms. Journal of Machine Learning Research, 23(274):1–18, 2022.
- The Malmo platform for artificial intelligence experimentation. In International Joint Conference on Artificial Intelligence (IJCAI) of 2016, volume 16, pages 4246–4247, 2016.
- The NetHack Learning Environment. Advances in Neural Information Processing Systems (NeurIPS) of 2020, 33:7671–7684, 2020.
- Revisiting the Arcade Learning Environment: Evaluation protocols and open problems for general agents. Journal of Artificial Intelligence Research, 61:523–562, 2018.
- Ray: A distributed framework for emerging AI applications. In USENIX Symposium on Operating Systems Design and Implementation (OSDI) of 2018, pages 561–577, 2018.
- Dreaming of many worlds: Learning contextual world models aids zero-shot generalization. arXiv preprint arXiv:2403.10967, 2024.
- Stable-Baselines3: Reliable reinforcement learning implementations. Journal of Machine Learning Research, 22(268):1–8, 2021.
- MiniHack the planet: A sandbox for open-ended reinforcement learning research. Datasets and Benchmarks Track of the 2021 NeurIPS, 2021.
- skrl: Modular and flexible library for reinforcement learning. Journal of Machine Learning Research, 24(254):1–9, 2023.
- MuJoCo: A physics engine for model-based control. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) of 2012, pages 5026–5033, 2012.
- ViZDoom Competitions: Playing Doom from Pixels. IEEE Transactions on Games, 11(3):248–259, 2019.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.