Search papers, labs, and topics across Lattice.
University of California
3
0
6
Restoring global logit comparability in vision-language models boosts zero-shot classification accuracy by 15% while maintaining robustness across diverse domains.
LLMs can achieve expert-level accuracy in specialized industrial domains by grounding them in traceable, queryable knowledge graphs built from fragmented scientific literature.
Synthesizing PBR materials from text and geometry just got a whole lot easier, thanks to a video diffusion transformer that jointly models material properties.
