CCL: Collaborative Curriculum Learning for Sparse-Reward Multi-Agent Reinforcement Learning via Co-evolutionary Task Evolution

2025/05/08 by Lin, Yufei, Ye, Chengwei, Zhang, Huanzhen +4 · 2 citations
#Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Multiagent Systems (cs.MA)

paper · doi:10.48550/arxiv.2505.07854

Abstract

Sparse reward environments pose significant challenges in reinforcement learning, especially within multi-agent systems (MAS) where feedback is delayed and shared across agents, leading to suboptimal learning. We propose Collaborative Multi-dimensional Course Learning (CCL), a novel curriculum learning framework that addresses this by (1) refining intermediate tasks for individual agents, (2) using a variational evolutionary algorithm to generate informative subtasks, and (3) co-evolving agents with their environment to enhance training stability. Experiments on five cooperative tasks in the MPE and Hide-and-Seek environments show that CCL outperforms existing methods in sparse reward settings.

Citations

• Efficient Knowledge Transfer in Multi-Task Learning through Task-Adaptive Low-Rank Representation
• CSCE: Boosting LLM Reasoning by Simultaneous Enhancing of Causal Significance and Consistency
• Synergistic Spotting and Recognition of Micro-Expression via Temporal State Transition
• Learning to Utilize Shaping Rewards: A New Approach of Reward Shaping
• Emergent Tool Use From Multi-Agent Autocurricula
• Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions
• Quantifying Generalization in Reinforcement Learning
• Reverse Curriculum Generation for Reinforcement Learning
• Attention Is All You Need
• Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
• Automatic Goal Generation for Reinforcement Learning Agents
• One-Shot Imitation Learning
• Safe, Multi-Agent, Reinforcement Learning for Autonomous Driving
• A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning

Cited by

• Graph-Enhanced Policy Optimization in LLM Agent Training
• GaMNet: A Hybrid Network with Gabor Fusion and NMamba for Efficient 3D Glioma Segmentation

Related

• Variational Automatic Curriculum Learning for Sparse-Reward Cooperative Multi-Agent Problems
• Towards Skilled Population Curriculum for Multi-Agent Reinforcement Learning
• Curriculum Learning With Counterfactual Group Relative Policy Advantage For Multi-Agent Reinforcement Learning
• Curriculum Learning for Cooperation in Multi-Agent Reinforcement Learning
• Learning Progress Driven Multi-Agent Curriculum