Search papers, labs, and topics across Lattice.
University of Stuttgart, IMPRS-IS * Equal Contributors
3
0
3
Selecting the right initial noise can dramatically enhance the coherence of robot action sequences without the need for retraining or policy changes.
EquiVLA achieves a remarkable 92.6% success rate in robotic tasks by leveraging rotational equivariance, setting a new standard for generalist manipulation models.
Advanced Vision-Language-Action models can be dramatically compressed by up to 50% without losing performance, reshaping our approach to robotic manipulation.
