Relational Pooling for Graph Representations

2019/03/06 by Ryan L. Murphy, Murphy, Ryan L., Balasubramaniam Srinivasan +5 · 16 citations
Computer Science · #Advanced Graph Neural Networks #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Stochastic Gradient Optimization Techniques #Topic Modeling

paper · pdf · doi:10.48550/arxiv.1903.02541

openalex publication_date 2019/03/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This work generalizes graph neural networks (GNNs) beyond those based on the Weisfeiler-Lehman (WL) algorithm, graph Laplacians, and diffusions. Our approach, denoted Relational Pooling (RP), draws from the theory of finite partial exchangeability to provide a framework with maximal representation power for graphs. RP can work with existing graph representation models and, somewhat counterintuitively, can make them even more powerful than the original WL isomorphism test. Additionally, RP allows architectures like Recurrent Neural Networks and Convolutional Neural Networks to be used in a theoretically sound approach for graph classification. We demonstrate improved performance of RP-based graph representations over state-of-the-art methods on a number of tasks.

Citations

• A Stochastic Approximation Method
• Backpropagation Applied to Handwritten Zip Code Recognition
• Molecular graph convolutions: moving beyond fingerprints
• Geometric Deep Learning: Going beyond Euclidean data
• Deep Models of Interactions Across Sets
• Detecting Functionality-Specific Vulnerabilities via Retrieving Individual Functionality-Equivalent APIs in Open-Source Repositories

Cited by

• CondPSE: A Polynomial-Filtered Structural Encoder with Conditional Modulation for Graphs
• Plain Transformers are Surprisingly Powerful Link Predictors
• Semantic Refinement with LLMs for Graph Representations
• Are Graph Transformers Necessary? Efficient Long-Range Message Passing with Fractal Nodes in MPNNs
• Sketch-Augmented Features Improve Learning Long-Range Dependencies in Graph Neural Networks
• Random Search Neural Networks for Efficient and Expressive Graph Learning
• Message Passing on the Edge: Towards Scalable and Expressive GNNs
• Exact Subgraph Isomorphism Network with Mixed L_0,2 Norm Constraint for Predictive Graph Mining
• Architectural change in neural networks using fuzzy vertex pooling
• Improving Expressivity of Graph Neural Networks
• A pipeline for fair comparison of graph neural networks in node classification tasks
• Resolving Indirect Calls in Binary Code via Cross-Reference Augmented Graph Neural Networks
• Principal Neighbourhood Aggregation for Graph Nets
• Breaking the Expressive Bottlenecks of Graph Neural Networks
• Bridging Theory and Practice in Link Representation with Graph Neural Networks
• Towards Scale-Invariant Graph-related Problem Solving by Iterative Homogeneous Graph Neural Networks

Related

• From Relational Pooling to Subgraph GNNs: A Universal Framework for More Expressive Graph Neural Networks
• Graph Pooling for Graph Neural Networks: Progress, Challenges, and Opportunities
• Accurate Learning of Graph Representations with Graph Multiset Pooling
• Weisfeiler and Leman Go Relational
• Clique pooling for graph classification