NTT's physics-based world model lifts a robot's shelf-stacking success from 73% to 92%, and Mujin will test it on airport baggage
NTT fitted a video-generation world model with a "physics control adapter" so a robot can preview whether a load will topple before it moves. Success on a small robot's shelf-stacking task rose from 73.1% to 92.3%, and field trials with Mujin start with airport baggage.
2 min readSources verified