A Parallel and Efficient Algorithm for Learning to Match

TL;DR

This paper introduces a novel parallel coordinate descent algorithm for feature-based matrix factorization, enabling scalable and efficient learning-to-match in large-scale data mining tasks without sacrificing accuracy.

Contribution

The paper presents a new parallel algorithm that relaxes the objective for efficient updates, allowing scalable learning-to-match with guaranteed convergence on large datasets.

Findings

Handles hundreds of millions of instances and features on a single machine

Significantly improves efficiency over baseline methods

Maintains accuracy while scaling to large datasets

Abstract

Many tasks in data mining and related fields can be formalized as matching between objects in two heterogeneous domains, including collaborative filtering, link prediction, image tagging, and web search. Machine learning techniques, referred to as learning-to-match in this paper, have been successfully applied to the problems. Among them, a class of state-of-the-art methods, named feature-based matrix factorization, formalize the task as an extension to matrix factorization by incorporating auxiliary features into the model. Unfortunately, making those algorithms scale to real world problems is challenging, and simple parallelization strategies fail due to the complex cross talking patterns between sub-tasks. In this paper, we tackle this challenge with a novel parallel and efficient algorithm for feature-based matrix factorization. Our algorithm, based on coordinate descent, can easily handle hundreds of millions of instances and features on a single machine. The key recipe of this algorithm is an iterative relaxation of the objective to facilitate parallel updates of parameters, with guaranteed convergence on minimizing the original objective function. Experimental results demonstrate that the proposed method is effective on a wide range of matching problems, with efficiency significantly improved upon the baselines while accuracy retained unchanged.