Higher-Order Factorization Machines

TL;DR

This paper introduces the first efficient algorithms for training higher-order factorization machines, enabling high-dimensional feature modeling with reduced computational costs and new shared-parameter variants for improved efficiency.

Contribution

It presents the first generic algorithms for arbitrary-order HOFMs and introduces shared-parameter variants that reduce model size and prediction time.

Findings

Effective training algorithms for higher-order FMs.

Shared-parameter variants maintain accuracy with less computation.

Demonstrated on four link prediction tasks.

Abstract

Factorization machines (FMs) are a supervised learning approach that can use second-order feature combinations even when the data is very high-dimensional. Unfortunately, despite increasing interest in FMs, there exists to date no efficient training algorithm for higher-order FMs (HOFMs). In this paper, we present the first generic yet efficient algorithms for training arbitrary-order HOFMs. We also present new variants of HOFMs with shared parameters, which greatly reduce model size and prediction times while maintaining similar accuracy. We demonstrate the proposed approaches on four different link prediction tasks.