Search papers, labs, and topics across Lattice.
1
0
3
Despite advances in vision-language models, reasoning across sparse, multi-view observations remains surprisingly unsolved, with current models barely outperforming random guessing on a new benchmark.
