Search papers, labs, and topics across Lattice.
1
0
3
Layout-preserving text beats pixel-level visual cues for structured data extraction from documents, according to a new multimodal benchmark.
