Search papers, labs, and topics across Lattice.
3
0
6
2
Kimera-VIO emerges as the only SLAM system capable of robustly tracking in low-light conditions, outperforming its peers despite the challenges posed by reduced visibility.
Autonomous evaluation can boost task success rates for GUI agents by over 12% when noise is effectively modeled and corrected.
Agentic RAG's potential in low-resource languages like Ukrainian hinges on better retrieval, as even sophisticated agentic reasoning can't overcome poor initial document identification.
