Search papers, labs, and topics across Lattice.
Affiliation:
3
0
4
Robots can now intelligently choose camera angles to answer questions by predicting the utility of unseen views, bypassing costly real-time rendering.
NavWAM turns visual foresight into executable robot actions, outperforming traditional planning methods in real-world navigation scenarios.
Despite advances in VLMs, agents struggle with active perception in 3D environments, revealing a significant gap in performance compared to humans.
