Search papers, labs, and topics across Lattice.
3
0
4
Distinct sparse features from intermediate layers reveal how neural networks misclassify out-of-distribution samples, leading to a new state-of-the-art detection method.
Achieving a 95.8% success rate in resolving self-collisions, PoseShield transforms how we handle human pose estimation under extreme articulations.
MAPL achieves expert-level locomotion performance without the burden of manual reward design by harnessing the power of language models to evaluate multiple objectives simultaneously.
