Search papers, labs, and topics across Lattice.
3
0
7
SpecLoR cuts through noise and drift in text-to-video generation, achieving remarkable motion coherence with minimal computational cost.
Homogeneous neural networks can be trained with provable generalization guarantees using significantly larger stepsizes than previously thought possible.
LLM benchmark accuracy jumps 10% when evaluated on a cleaned-up version of Humanity's Last Exam, highlighting the significant impact of dataset noise on performance metrics.
