Interagent beliefs for learning to communicate in large-scale multirobot visual object searchPuc, Jernej (Avtor) Škulj, Gašper (Avtor) Pleterski, Jan (Avtor) Podržaj, Primož (Avtor) Vrabič, Rok (Avtor) cooperating robotsmultirobot systemsreinforcement learningvision-based navigationWe present Century Maze, a simulated environment for cooperative navigation of groups numbering over 100 robots with primarily visual observations. The underlying task of large-scale object search is designed to push the scalability and efficiency of decentralized multirobot systems relying on communication to overcome the limitations of separate units. To enable the training of robotic agents in large-scale visual embodied applications such as Century Maze, we propose two reinforcement learning (RL) components that facilitate the emergence of effective communication: Differentiable Interagent Belief Learning (DIABL), which provides a clear learning signal by optimizing the agents' beliefs about their goals and propagating gradients back over a differentiable communication channel to include the appropriate observers, and informative event replay (infer), which compensates for the scarcity of samples relevant to DIABL by maintaining a dedicated buffer for auxiliary learning. We demonstrate our combined method through an actor-critic neural network architecture encompassing a decentralized multiagent system of variable size with individual goal-based belief modules. Results on Century Maze show that our method is able to significantly improve performance through cooperation over 100 agents against comparable baselines and maintains its effectiveness in realistic conditions.20262026-05-12 12:27:42Članek v reviji182457UDK: 007.52ISSN pri članku: 1941-0468DOI: 10.1109/TRO.2026.3677944COBISS_ID: 277860355sl