Search papers, labs, and topics across Lattice.
3
0
3
Achieving stable grasping of both soft and rigid objects using only visual input could revolutionize robotic manipulation in everyday settings.
LOPAL achieves a remarkable 27.31% boost in task performance by intelligently leveraging local quality assessments from imperfect human demonstrations.
Vision-free robotic manipulation policies can be trained in under an hour to achieve 95% success and generalize to unseen tasks by distilling knowledge from vision-enabled teachers using real-world reinforcement learning.
