Search papers, labs, and topics across Lattice.
1
0
2
Location-aware pretraining unlocks significant gains in medical difference VQA by teaching models to spot clinically relevant changes that standard vision encoders miss.
