Imitation Bootstrapped Multi-Agent Reinforcement Learning
arXiv cs.AI · 14 Nov 2026, 04:50 AEST
Imitation Bootstrapped Multi-Agent Reinforcement Learning (IBMARL) enhances multi-agent reinforcement learning by utilizing a framework that incorporates behavior cloning. This method allows agents to learn from both imitation and reinforcement learning policies, significantly improving performance on cooperative tasks.
Key facts
- Demonstrations are an effective way to accelerate single-agent reinforcement learning (RL), but extending this idea to multi-agent reinforcement learning (MARL) is non-trivial as it is difficult and often impossible for a single person to teleoperate multiple agents simultaneously.
- Round-Robin Behavior Cloning (R2BC) addresses this by letting a single demonstrator control one agent at a time while the others execute their current policies.
- We propose Imitation Bootstrapped Multi-Agent Reinforcement Learning (IBMARL), a framework that extends Imitation Bootstrapped Reinforcement Learning (IBRL) to cooperative teams and uses R2BC to train the imitation policies.
- On four cooperative multi-agent RL benchmarks, IBMARL surpasses its imitation policies $2.6$-$4.2 imes$ faster than demonstration-based baselines on two tasks, is the only method to surpass it on the other two, and exceeds the demonstrator itself on two of the four.
Read aloud
Imitation Bootstrapped Multi-Agent Reinforcement Learning IBMARL enhances multi-agent reinforcement learning by utilizing a framework that incorporates behavior cloning. This method allows agents to learn from both imitation and reinforcement learning policies, significantly improving performance on cooperative tasks. Demonstrations are an effective way to accelerate single-agent reinforcement learning RL, but extending this idea to multi-agent reinforcement learning MARL is non-trivial as it is difficult and often impossible for a single person to teleoperate multiple agents simultaneously. Round-Robin Behavior Cloning R2BC addresses this by letting a single demonstrator control one agent at a time while the others execute their current policies. We propose Imitation Bootstrapped Multi-Agent Reinforcement Learning IBMARL, a framework that extends Imitation Bootstrapped Reinforcement Learning IBRL to cooperative teams and uses R2BC to train the imitation policies. On four cooperative multi-agent RL benchmarks, IBMARL surpasses its imitation policies two point six dollars$-four point two dollars imes$ faster than demonstration-based baselines on two tasks, is the only method to surpass it on the other two, and exceeds the demonstrator itself on two of the four.
Sources in this cluster
- Imitation Bootstrapped Multi-Agent Reinforcement Learning
arXiv cs.AI · 14 Nov 2026, 04:50 AEST