Logistic Tensor Factorization for Multi-Relational Data

TL;DR

This paper introduces a logistic extension to the RESCAL tensor factorization method, improving multi-relational data prediction by accounting for the binary nature of adjacency tensors, with significant performance gains on benchmark datasets.

Contribution

It presents a novel logistic tensor factorization method that enhances RESCAL for binary multi-relational data, demonstrating improved predictive accuracy.

Findings

Logistic extension significantly improves prediction results.

Enhanced RESCAL outperforms previous methods on benchmarks.

Binary-aware tensor factorization yields better modeling of adjacency data.

Abstract

Tensor factorizations have become increasingly popular approaches for various learning tasks on structured data. In this work, we extend the RESCAL tensor factorization, which has shown state-of-the-art results for multi-relational learning, to account for the binary nature of adjacency tensors. We study the improvements that can be gained via this approach on various benchmark datasets and show that the logistic extension can improve the prediction results significantly.