SEP 10, 2026 · PREPRINT
Reification as a Transferable Vocabulary: Zero-Shot Link Prediction with Vanilla GNNs
arXiv
A proof-of-concept study demonstrating that a representational approach (reification) enables zero-shot transfer in graph neural networks, with promising results on benchmarks but limited scope, no peer review, and early-stage database evaluation.
PreliminaryCamille Pradel
Reported
Pretraining knowledge graph size4,245 triples
Number of transfer benchmarks40 inductive link-prediction benchmarks
Training time per model30 minutes on one NVIDIA A100