Search papers, labs, and topics across Lattice.
5
0
8
V2V-Bench reveals that leading models excel in different aspects of video generation, highlighting the nuanced strengths of Grok and Veo3 in editing fidelity and visual quality, respectively.
LongSpace reveals that integrating explicit spatial memory into MLLMs can dramatically enhance their performance on long-horizon tasks, a critical advancement for applications in autonomous navigation.
Modeling foot sinkage and slip with three extra degrees of freedom unlocks more accurate bipedal robot locomotion on challenging granular terrains.
Optimizing federated learning incentives and privacy requires considering the interplay of data volume and noise level, not just one in isolation.
Ditch the paired data and deep learning: this unsupervised video stabilization method rivals offline techniques while running online, even on resource-constrained hardware.
