2000 character limit reached
CompetEvo: Towards Morphological Evolution from Competition (2405.18300v1)
Published 28 May 2024 in cs.AI
Abstract: Training an agent to adapt to specific tasks through co-optimization of morphology and control has widely attracted attention. However, whether there exists an optimal configuration and tactics for agents in a multiagent competition scenario is still an issue that is challenging to definitively conclude. In this context, we propose competitive evolution (CompetEvo), which co-evolves agents' designs and tactics in confrontation. We build arenas consisting of three animals and their evolved derivatives, placing agents with different morphologies in direct competition with each other. The results reveal that our method enables agents to evolve a more suitable design and strategy for fighting compared to fixed-morph agents, allowing them to obtain advantages in combat scenarios. Moreover, we demonstrate the amazing and impressive behaviors that emerge when confrontations are conducted under asymmetrical morphs.
- Continuous adaptation via meta-learning in nonstationary and competitive environments. 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings, (March):1–21, 2018.
- Emergent complexity via multi-agent competition. 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings, 2018.
- Task2Morph: Differentiable Task-Inspired Framework for Contact-Aware Robot Design. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 452–459, 2023.
- Evolving Physical Instinct for Morphology and Control Co-Adaption. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 6616–6623, 2023.
- Cooperative distributed model predictive control for robot in-hand manipulation. Robotic Intelligence and Automation, 43(1):65–74, 2023.
- Adversarial Policies: Attacking Deep Reinforcement Learning. 8th International Conference on Learning Representations, ICLR 2020, 2020.
- Embodied intelligence via learning and evolution. Nature Communications, 12(1), 2021.
- David Ha. Reinforcement learning for improving agent design. Artificial Life, 25(4):352–365, 2019.
- Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning. arXiv preprint arXiv:2304.13653, 2023.
- GLSO: Grammar-guided Latent Space Optimization for Sample-efficient Robot Design Automation. Proceedings of Machine Learning Research, 205:1321–1331, 2023.
- Decentralised aerial swarm for adaptive and energy efficient transport of unknown loads. Swarm and Evolutionary Computation, 67, 2021.
- Bio-inspired Multi-agent Model and Optimization Strategy for Collaborative Aerial Transport. Lecture Notes in Electrical Engineering, 801 LNEE:591–598, 2022.
- Stimulate the Potential of Robots via Competition. arXiv preprint arXiv:2403.10487, 2024.
- Emergent coordination through competition. 7th International Conference on Learning Representations, ICLR 2019, 2019.
- From motor control to team play in simulated humanoid football. Science Robotics, 7(69), 2022.
- Morphology-based embodied intelligence: Historical retrospect and research progress. Acta Autom. Sinica, 49(6):1131–1154, 2023.
- Learning by playing - Solving sparse reward tasks from scratch. 35th International Conference on Machine Learning, ICML 2018, 10:6910–6919, 2018.
- Jointly learning to construct and control agents using deep reinforcement learning. Proceedings - IEEE International Conference on Robotics and Automation, 2019-May:9798–9805, 2019.
- Mastering the game of Go with deep neural networks and tree search. Nature, 529(7587):484–489, 2016.
- A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play. Science, 362(6419):1140–1144, 2018.
- Karl Sims. Evolving Virtual Creatures. Seminal Graphics Papers: Pushing the Boundaries, Volume 2, pages 699–706, 2023.
- Grandmaster level in StarCraft II using multi-agent reinforcement learning. Nature, 575(7782):350–354, 2019.
- Nervenet: Learning structured policy with graph neural networks. 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings, 2018.
- Neural graph evolution: Towards efficient automatic robot design. 7th International Conference on Learning Representations, ICLR 2019, 2019.
- Multi-robot raster map fusion without initial relative position. Robotic Intelligence and Automation, 43(5):498–508, 2023.
- PreCo: Enhancing Generalization in Co-Design of Modular Soft Robots via Brain-Body Pre-Training. Proceedings of Machine Learning Research, 229:478–498, 2023.
- Transform2Act: Learning a Transform-and-Control Policy for Efficient Agent Design. ICLR 2022 - 10th International Conference on Learning Representations, 2022.
- Robust Multi-Agent Coordination via Evolutionary Generation of Auxiliary Adversarial Attackers. Proceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI 2023, 37:11753–11762, 2023.
- A Multi-modal Deformable Land-air Robot for Complex Environments. 2023 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 2022.
- Coupled Modeling and Fusion Control for a Multi-modal Deformable Land-air Robot. arXiv preprint arXiv:2211.04185, 2022.