Search papers, labs, and topics across Lattice.
Affiliation:
1
0
0
Grounding VLA policies with predictive, whole-body tactile representations boosts real-world humanoid loco-manipulation success by 43 percentage points over standard vision-proprioception approaches.
