Search papers, labs, and topics across Lattice.
1
0
2
Query-conditioned visual evidence graphs can boost multimodal reasoning accuracy by over 20% while using significantly less image data.
