Search papers, labs, and topics across Lattice.
2
0
6
3
ACE-Data-0 reveals that existing models struggle significantly with complex interactions, exposing critical gaps in embodied AI performance.
Current video generation models struggle with law-grounded reasoning, with the best achieving only 47% on the new Apple-PI benchmark.
