Search papers, labs, and topics across Lattice.
Karlsruhe Institute of Technology
6
0
6
INSPECT is introduced, which learns robot view preferences from records of a smart-glasses assistant that answers part queries and provides next-step guidance and achieves the highest view utility among the compared non-oracle policies and raises human-rated full verifiability.
MS-MEM achieves higher mapping accuracy while minimizing scene disturbances, showcasing the power of integrating multiple manipulation skills in robotic perception.
Aligning utility training with policy optimization, PAWS overcomes distribution shifts that cripple traditional preference-based reinforcement learning methods.
HOWTransfer achieves an 86% success rate in translating human hand motions into robot actions, surpassing teleoperation in user preference.
Decomposing pose estimation into error attribution and targeted mitigation allows simple algorithms like ICP to rival the robustness of complex foundation models, but with significantly less computation.
Rescue your robot's near-miss manipulation failures with FlowCorrect, a VR-guided, human-in-the-loop correction framework that boosts success rates by 85% without full retraining.
