GeniePath: Graph Neural Networks with Adaptive Receptive Paths

Abstract

We present, GeniePath, a scalable approach for learning adaptive receptive fields of neural networks defined on permutation invariant graph data. In GeniePath, we propose an adaptive path layer consists of two complementary functions designed for breadth and depth exploration respectively, where the former learns the importance of different sized neighborhoods, while the latter extracts and filters signals aggregated from neighbors of different hops away. Our method works in both transductive and inductive settings, and extensive experiments compared with competitive methods show that our approaches yield state-of-the-art results on large graphs.

References in corpus (3)

• Semi-Supervised Classification with Graph Convolutional Networks
• Inductive Representation Learning on Large Graphs
• Predicting multicellular function through multi-layer tissue networks

Cited by in corpus (13)

• GraphNAS: Graph Neural Architecture Search with Reinforcement Learning
• Janossy Pooling: Learning Deep Permutation-Invariant Functions for Variable-Size Inputs
• Uncovering Insurance Fraud Conspiracy with Network Learning
• xFraud: Explainable Fraud Transaction Detection
• Multi-hop Attention Graph Neural Network
• Improving Graph Attention Networks with Large Margin-based Constraints
• InceptionGCN: Receptive Field Aware Graph Convolutional Network for Disease Prediction
• SPAGAN: Shortest Path Graph Attention Network
• Heterogeneous Graph Neural Networks for Malicious Account Detection
• InfDetect: a Large Scale Graph-based Fraud Detection System for E-Commerce Insurance
• Graph Attention Recurrent Neural Networks for Correlated Time Series Forecasting -- Full version
• GAIN: Graph Attention & Interaction Network for Inductive Semi-Supervised Learning over Large-scale Graphs
• Graph Representation Learning for Merchant Incentive Optimization in Mobile Payment Marketing