How Many and How Much? Contrastive Learning with Negative Samples for Knowledge Graph Embedding
Asks how many negatives and how hard they should be for contrastive knowledge graph embedding.
Explainer · on the wire
The explainer for this story is still on the wire.
The claim, the mechanism, and the figures land here once generated.
Source · arXiv wire