Influence-Based Reinforcement Learning for Intrinsically-Motivated Agents
This addresses the problem of exponential joint action space growth in MARL for researchers and practitioners, though it appears incremental as it builds on existing intrinsic motivation concepts.
The paper tackles the challenge of achieving sufficient exploration and coordination in multi-agent reinforcement learning by rewarding agents for contributing to diversified team behavior through intrinsic motivation functions, resulting in significantly improved performances on tasks like the StarCraft Multi-Agent Challenge.
Discovering successful coordinated behaviors is a central challenge in Multi-Agent Reinforcement Learning (MARL) since it requires exploring a joint action space that grows exponentially with the number of agents. In this paper, we propose a mechanism for achieving sufficient exploration and coordination in a team of agents. Specifically, agents are rewarded for contributing to a more diversified team behavior by employing proper intrinsic motivation functions. To learn meaningful coordination protocols, we structure agents' interactions by introducing a novel framework, where at each timestep, an agent simulates counterfactual rollouts of its policy and, through a sequence of computations, assesses the gap between other agents' current behaviors and their targets. Actions that minimize the gap are considered highly influential and are rewarded. We evaluate our approach on a set of challenging tasks with sparse rewards and partial observability that require learning complex cooperative strategies under a proper exploration scheme, such as the StarCraft Multi-Agent Challenge. Our methods show significantly improved performances over different baselines across all tasks.