CORE MACHINE LEARNING

Revisiting Graph Neural Networks for Link Prediction

December 06, 2021

Abstract

In this paper, we theoretically characterize graph neural network’s representation power for high-order node set prediction problems (where a prediction is made over a set of more than 1 node). In particular, we focus on one most important second-order task—link prediction. There are two representative classes of GNN methods for link prediction: GAE and SEAL. GAE (Graph Autoencoder) first applies a GNN to the whole graph, and then aggregates the representations of the source and target nodes as their link representation. SEAL extracts a subgraph around the source and target nodes, labels the nodes in the subgraph, and then uses a GNN to learn a link representation from the labeled subgraph. At first glance, both GAE and SEAL use a GNN. However, their performance gap can be very large. On the recent Open Graph Benchmark datasets, SEAL achieved 3 first places out of 4 datasets, outperforming the best GAE method by up to 195% in Hits@100. In this paper, by studying this performance gap between GAE and SEAL, we first point out a key limitation of GAE caused by directly aggregating two node representations as a link representation. To address this limitation, we propose the labeling trick. Labeling trick unifies several recent successes to improve GNNs’ representation power, such as SEAL, Distance Encoding, and Identity-aware GNN, into a single and most basic form. We prove that with labeling trick a sufficiently expressive GNN can learn the most expressive structural representations for node sets. Our work establishes a theoretical foundation for using GNNs for high-order node set prediction.

AUTHORS

Written by

Yinglong Xia

Publisher

NeurIPS

Research Topics

Core Machine Learning

Related Publications

December 06, 2021

COMPUTER VISION

CORE MACHINE LEARNING

Debugging the Internals of Convolutional Networks

Bilal Alsallakh, Narine Kokhlikyan, Vivek Miglani, Shubham Muttepawar, Edward Wang (AI Infra), Sara Zhang, David Adkins, Orion Reblitz-Richardson

December 06, 2021

December 06, 2021

INTEGRITY

CORE MACHINE LEARNING

BulletTrain: Accelerating Robust Neural Network Training via Boundary Example Mining

Weizhe Hua, Yichi Zhang, Chuan Guo, Zhiru Zhang, Edward Suh

December 06, 2021

December 06, 2021

THEORY

CORE MACHINE LEARNING

Learning on Random Balls is Sufficient for Estimating (Some) Graph Parameters

Takanori Maehara, Hoang NT

December 06, 2021

November 12, 2021

THEORY

REINFORCEMENT LEARNING

Bandits with Knapsacks beyond the Worst-Case Analysis

Karthik Abinav Sankararaman, Aleksandrs Slivkins

November 12, 2021