Explore SPACeR, a novel framework merging self-play reinforcement learning with centralized reference models to achieve efficient, socially aware autonomous ...
Level: advanced
By Wei-Jer Chang and 6 other authors
Category: research