Taiki Miyanishi

Robots can now intelligently choose camera angles to answer questions by predicting the utility of unseen views, bypassing costly real-time rendering.

Existing Video REC models falter dramatically when faced with the complexities of long-form egocentric videos, revealing a critical gap in current methodologies.

NavWAM turns visual foresight into executable robot actions, outperforming traditional planning methods in real-world navigation scenarios.