T-PAIR: Temporal Node-Pair Embedding for Automatic Biomedical Hypothesis Generation

Abstract

T-PAIR represents node pairs at specific time points in a biomedical knowledge graph. By learning how relationships evolve over time, the model can predict which connections are likely to become established — effectively generating ranked candidate hypotheses for biomedical research.

One of my most-cited papers (32 citations) — a core contribution to temporal hypothesis generation.