Jul 2026
GraphRareBench: An Auditable Graph-Evidence Benchmark for Phenotype-Driven Rare-Disease Diagnosis
Results indicate that full-pool retrieval, hard-confounder discrimination, and observable evidence access capture complementary aspects of model behavior, which provides a foundation for more transparent and evidence-aware evaluation of phenotype-driven diagnostic systems.
Guiling Guo, Jia Yang, Jia-Hao Xu et al.
· arXiv.org · 0 citations