Yimiao Feng

This paper introduces LukePi, a self-supervised pre-training framework for GNNs on biomedical knowledge graphs, which improves link prediction in low-data and distribution-shift scenarios.

MiT4SL predicts synthetic lethality in cancer cell lines using multi-omics triplet representation learning, outperforming existing models in cell line-adapted scenarios.

This study benchmarks machine learning methods for predicting synthetic lethality in cancer using matrix factorization and deep learning models, highlighting performance differences and emphasizing data and scenario importance.