MixHop: Higher-Order Graph Convolutional Architectures via Sparsified\n Neighborhood Mixing

2019/04/30 by Sami Abu-El-Haija, Bryan Perozzi, Abu-El-Haija, Sami +13 · 29 citations
Computer Science · Health Professions · #Advanced Graph Neural Networks #Artificial Intelligence in Healthcare #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning in Healthcare #Social and Information Networks (cs.SI)

paper · pdf · doi:10.48550/arxiv.1905.00067

openalex publication_date 2019/04/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Existing popular methods for semi-supervised learning with Graph Neural\nNetworks (such as the Graph Convolutional Network) provably cannot learn a\ngeneral class of neighborhood mixing relationships. To address this weakness,\nwe propose a new model, MixHop, that can learn these relationships, including\ndifference operators, by repeatedly mixing feature representations of neighbors\nat various distances. Mixhop requires no additional memory or computational\ncomplexity, and outperforms on challenging baselines. In addition, we propose\nsparsity regularization that allows us to visualize how the network prioritizes\nneighborhood information across different graph datasets. Our analysis of the\nlearned architectures reveals that neighborhood mixing varies per datasets.\n

Citations

• ImageNet classification with deep convolutional neural networks
• Detecting Functionality-Specific Vulnerabilities via Retrieving Individual Functionality-Equivalent APIs in Open-Source Repositories

Cited by

• Task-driven Heterophilic Graph Structure Learning
• Grad: Guided Relation Diffusion Generation for Graph Augmentation in Graph Fraud Detection
• Can You Hear Me Now? A Benchmark for Long-Range Graph Propagation
• ATLAS: Adaptive Topology-based Learning at Scale for Homophilic and Heterophilic Graphs
• Polynomial Neural Sheaf Diffusion: A Spectral Filtering Approach on Cellular Sheaves
• Hypergraph Contrastive Learning for both Homophilic and Heterophilic Hypergraphs
• Gauge-Equivariant Graph Networks via Self-Interference Cancellation
• Complex-Weighted Convolutional Networks: Provable Expressiveness via Complex Diffusion
• AutoHood3D: A Multi-Modal Benchmark for Automotive Hood Design and Fluid-Structure Interaction
• Fixed-point graph convolutional networks against adversarial attacks
• An Active Diffusion Neural Network for Graphs
• Training Diverse Graph Experts for Ensembles: A Systematic Empirical Study
• Fighter: Unveiling the Graph Convolutional Nature of Transformers in Time Series Modeling
• A Hierarchical Quantized Tokenization Framework for Task-Adaptive Graph Representation Learning
• Graph Few-Shot Learning via Adaptive Spectrum Experts and Cross-Set Distribution Calibration
• Multi-View Graph Feature Propagation for Privacy Preservation and Feature Sparsity
• Glance for Context: Learning When to Leverage LLMs for Node-Aware GNN-LLM Fusion
• Soft-Evidence Fused Graph Neural Network for Cancer Driver Gene Identification across Multi-View Biological Graphs
• Exact Subgraph Isomorphism Network with Mixed L_0,2 Norm Constraint for Predictive Graph Mining
• Bridging Computational Social Science and Deep Learning: Cultural Dissemination-Inspired Graph Neural Networks
• Asynchronous Message Passing for Addressing Oversquashing in Graph Neural Networks
• Unifying Adversarial Perturbation for Graph Neural Networks
• Memorization in Graph Neural Networks
• Training Transformers for Mesh-Based Simulations
• Addressing Graph Anomaly Detection via Causal Edge Separation and Spectrum
• Over-Squashing in GNNs and Causal Inference of Rewiring Strategies
• TANGO: Graph Neural Dynamics via Learned Energy and Tangential Flows
• Global Context Enhanced Social Recommendation with Hierarchical Graph Neural Networks
• Mutual Information Maximization in Graph Neural Networks

Related

• On the Equivalence of Graph Convolution and Mixup
• G-Mixup: Graph Data Augmentation for Graph Classification
• Neighborhood Convolutional Network: A New Paradigm of Graph Neural Networks for Node Classification
• Semi-supervised Learning with Adaptive Neighborhood Graph Propagation Network
• Teaching Yourself: Graph Self-Distillation on Neighborhood for Node Classification