Graph Convolutional Matrix Completion

TL;DR

This paper introduces a graph auto-encoder framework for matrix completion in recommender systems, leveraging deep learning on bipartite graphs to improve prediction accuracy especially when additional structured data is available.

Contribution

It proposes a novel graph auto-encoder approach using differentiable message passing for matrix completion, outperforming existing methods with extra feature information.

Findings

Competitive performance on standard benchmarks

Outperforms state-of-the-art with additional structured data

Effective in link prediction on bipartite graphs

Abstract

We consider matrix completion for recommender systems from the point of view of link prediction on graphs. Interaction data such as movie ratings can be represented by a bipartite user-item graph with labeled edges denoting observed ratings. Building on recent progress in deep learning on graph-structured data, we propose a graph auto-encoder framework based on differentiable message passing on the bipartite interaction graph. Our model shows competitive performance on standard collaborative filtering benchmarks. In settings where complimentary feature information or structured data such as a social network is available, our framework outperforms recent state-of-the-art methods.