Learning Invariant Graph Representations for Cox Survival Modeling under Distribution Shifts
Karaseva Ng, Chengshang Lyu, Anna Jiang, Yinhu Li, Lingxi Chen
This paper introduces InvGraphCox, a framework that improves survival prediction by learning invariant graph representations under distribution shifts, outperforming baselines in discrimination and risk stratification.
bioRxiv · Dec 12, 2025 · PreprintBioinformatics