Adaptive Coordination in Social Embodied Rearrangement
This addresses the challenge of robots adapting to new human partners in everyday cooperative tasks, representing an incremental improvement in multi-agent coordination.
The paper tackled the problem of zero-shot coordination in complex, visually rich multi-agent tasks like setting a dinner table, proposing Behavior Diversity Play (BDP) to encourage diverse behaviors during training, which resulted in 35% higher success and 32% higher efficiency compared to baselines.
We present the task of "Social Rearrangement", consisting of cooperative everyday tasks like setting up the dinner table, tidying a house or unpacking groceries in a simulated multi-agent environment. In Social Rearrangement, two robots coordinate to complete a long-horizon task, using onboard sensing and egocentric observations, and no privileged information about the environment. We study zero-shot coordination (ZSC) in this task, where an agent collaborates with a new partner, emulating a scenario where a robot collaborates with a new human partner. Prior ZSC approaches struggle to generalize in our complex and visually rich setting, and on further analysis, we find that they fail to generate diverse coordination behaviors at training time. To counter this, we propose Behavior Diversity Play (BDP), a novel ZSC approach that encourages diversity through a discriminability objective. Our results demonstrate that BDP learns adaptive agents that can tackle visual coordination, and zero-shot generalize to new partners in unseen environments, achieving 35% higher success and 32% higher efficiency compared to baselines.