A shallow neural model for relation prediction

TL;DR

This paper introduces SHALLOM, a shallow neural network model that predicts missing relations in knowledge graphs through multi-label classification, outperforming state-of-the-art methods with fast training times.

Contribution

The paper presents a novel shallow neural model for relation prediction in knowledge graphs, framing it as a multi-label classification task, with improved accuracy and efficiency.

Findings

SHALLOM outperforms existing models on FB15K-237 and WN18RR datasets.

SHALLOM achieves up to 8% absolute improvement in accuracy.

Training time is under 8 minutes on key datasets.

Abstract

Knowledge graph completion refers to predicting missing triples. Most approaches achieve this goal by predicting entities, given an entity and a relation. We predict missing triples via the relation prediction. To this end, we frame the relation prediction problem as a multi-label classification problem and propose a shallow neural model (SHALLOM) that accurately infers missing relations from entities. SHALLOM is analogous to C-BOW as both approaches predict a central token (p) given surrounding tokens ((s,o)). Our experiments indicate that SHALLOM outperforms state-of-the-art approaches on the FB15K-237 and WN18RR with margins of up to $3\%$ and $8\%$ (absolute), respectively, while requiring a maximum training time of 8 minutes on these datasets. We ensure the reproducibility of our results by providing an open-source implementation including training and evaluation scripts at {\url{https://github.com/dice-group/Shallom}.}