Ikuro Sato

Test-time planning for robots can be dramatically improved by focusing on selecting reliable future-action hypotheses rather than merely generating more of them.

Winning noise tickets can dramatically enhance text-to-motion fidelity, bridging the gap between input semantics and generated motion without the need for retraining.

Steal accuracy from dense models and stabilize MoE training with a simple teacher-guided routing scheme that combats gradient starvation.