Track trials, findings, and clinical evidence as it lands.
Search papers, labs, and topics across Lattice.
3
0
4
1
Local pretraining objectives don't just dominate fine-grained pathology detection—they unexpectedly rival global contrastive models on high-level report generation and QA, upending assumptions about cross-granularity transfer.
Compact, fine-tuned ASR models can outperform massive multilingual models by 26.9 percentage points in recognizing spontaneous African speech.
Location-aware pretraining unlocks significant gains in medical difference VQA by teaching models to spot clinically relevant changes that standard vision encoders miss.