SimWAM: A Simple World Action Model for End-to-End Autonomous Driving
The episode discusses SimWAM, a World Action Model for autonomous driving that uses training signals from future video prediction to teach an action expert, avoiding costly test-time future imagination. Hosts discuss how joint flow matching links video and action experts, leading to fast inference and modular scalability.