Distmult model implementation for

This project is a Distmult model implementation for Knowledge graph completion

Details of implementation

Sample corruption: Random replacement of head/tail
Dataset choice: FB15K-237
Loss function: Multiclass Negative Log-likelihood
Tie-breaking Policy: random

[TODO: ]Extension

Sample corruption: Bernouli distribution, Self-adversarial
Tie-breaking Policy: min/max
Loss function: Pairwise Ranking