Search papers, labs, and topics across Lattice.
4
52
5
12
A single 500-hour visual-tactile dataset boosts zero-shot contact prediction from 13% to 38% IoU and quadruples robot manipulation success rates, proving scale and consistency are paramount for embodied learning.
Throwing away learned attention in favor of pure 3D spatial coverage allows VLMs to retain over 93% of their multi-view 3D reasoning performance while slashing visual token budgets by ~92%.
Generating missing multi-view data from diverse driving videos boosts closed-loop driving robustness in edge cases by over 30%.
Unlock human-like spatial reasoning in VLMs with VLM-3R, which reconstructs 3D understanding from monocular video using instruction tuning, bypassing the need for external depth sensors.
