Search papers, labs, and topics across Lattice.
3
0
4
3
Equal nominal success in dexterous manipulation doesn't mean your AI can handle speed variations like a human expert.
FIRMGrasp reveals that over half of the grasps deemed high-quality by traditional metrics fail under adverse friction conditions, highlighting a critical gap in grasp assessment.
Variational neural belief parameterizations can drastically enhance grasping success rates and reduce planning time under multimodal uncertainties, outperforming traditional methods by an order of magnitude.
