# TuckER: Tensor Factorization for Knowledge Graph Completion

## TL;DR

TuckER is a linear tensor factorization model for knowledge graph completion that outperforms previous models and serves as a strong baseline for link prediction tasks.

## Contribution

The paper introduces TuckER, a Tucker decomposition-based model that is fully expressive and unifies several existing linear models for knowledge graph completion.

## Key findings

- TuckER outperforms previous state-of-the-art models on standard datasets.
- TuckER is fully expressive with derived bounds on embedding dimensions.
- Several existing linear models are special cases of TuckER.

## Abstract

Knowledge graphs are structured representations of real world facts. However, they typically contain only a small subset of all possible facts. Link prediction is a task of inferring missing facts based on existing ones. We propose TuckER, a relatively straightforward but powerful linear model based on Tucker decomposition of the binary tensor representation of knowledge graph triples. TuckER outperforms previous state-of-the-art models across standard link prediction datasets, acting as a strong baseline for more elaborate models. We show that TuckER is a fully expressive model, derive sufficient bounds on its embedding dimensionalities and demonstrate that several previously introduced linear models can be viewed as special cases of TuckER.