Search papers, labs, and topics across Lattice.
6
0
8
5
Achieving 74.8% accuracy on a new temporal reasoning benchmark, ChronoVision redefines how multimodal models can tackle complex visual tasks.
Current models struggle with the complexities of children's gait, revealing a critical gap in automated assessment tools for pediatric neuromuscular disorders.
Claw-like agents are vulnerable to severe security breaches, with malicious plugins achieving a 100% success rate in attacks.
A single generalist model outperforms specialized systems, achieving over 35% improvement in real-world robotic task success.
Cosmos 3 sets a new benchmark for omnimodal models, outperforming existing state-of-the-art in both Text-to-Image and Image-to-Video tasks.
Models are substantially better at pairwise self-verification than independent scoring, unlocking a more efficient and accurate approach to test-time scaling for complex reasoning.
