AI RESEARCH: PROXIMAL STATE NUDGING CUTS SKILL ATROPHY BY 7X IN AI-ASSISTED DRIVING

New shared autonomy algorithm from Stanford et al. nudges users toward learnable states. In CARLA simulator, PSN produced up to 7x unassisted skill gains and 50% fewer collisions vs. standard baselines.

New shared autonomy algorithm from Stanford et al. nudges users toward learnable states. In CARLA simulator, PSN produced up to 7x unassisted skill gains and 50% fewer collisions vs. standard baselines.