Search papers, labs, and topics across Lattice.
2
0
3
0
Multi-view reasoning can be pretrained and reused, enabling competitive performance without the need for constant retraining on downstream tasks.
DeepBD outperforms traditional variant prioritization tools by integrating LLMs and specialized evidence modules, achieving a Recall@10 of 92.9% on a large cohort of genetic cases.
