Hyperbolic Graph Neural Networks

Abstract

Learning from graph-structured data is an important task in machine learning and artificial intelligence, for which Graph Neural Networks (GNNs) have shown great promise. Motivated by recent advances in geometric representation learning, we propose a novel GNN architecture for learning representations on Riemannian manifolds with differentiable exponential and logarithmic maps. We develop a scalable algorithm for modeling the structural properties of graphs, comparing Euclidean and hyperbolic geometry. In our experiments, we show that hyperbolic GNNs can lead to substantial improvements on various benchmark datasets.

Published at NeurIPS 2019

Cited by in corpus (38)

• Hyperbolic Graph Convolutional Neural Networks
• Network Geometry
• Graph Convolutional Networks for Graphs Containing Missing Features
• Machine Learning on Graphs: A Model and Comprehensive Taxonomy
• Revisiting Over-smoothing in Deep GCNs
• Geoopt: Riemannian Optimization in PyTorch
• Graph-based Modeling of Online Communities for Fake News Detection
• ACE-HGNN: Adaptive Curvature Exploration Hyperbolic Graph Neural Network
• Constant Curvature Graph Convolutional Networks
• Understanding over-squashing and bottlenecks on graphs via curvature
• From Trees to Continuous Embeddings and Back: Hyperbolic Hierarchical Clustering
• Optimal Transport Graph Neural Networks
• Graph Geometry Interaction Learning
• Dorylus: Affordable, Scalable, and Accurate GNN Training with Distributed CPU Servers and Serverless Threads
• Hyperbolic Graph Attention Network
• Hyperbolic Deep Neural Networks: A Survey
• Knowledge Association with Hyperbolic Knowledge Graph Embeddings
• Coordinate Independent Convolutional Networks -- Isometry and Gauge Equivariant Convolutions on Riemannian Manifolds
• Latent Variable Modelling with Hyperbolic Normalizing Flows
• Size-Invariant Graph Representations for Graph Classification Extrapolations
• Hyperbolic Variational Graph Neural Network for Modeling Dynamic Graphs
• HCGR: Hyperbolic Contrastive Graph Representation Learning for Session-based Recommendation
• Privacy-Preserving Representation Learning on Graphs: A Mutual Information Perspective
• A Hyperbolic-to-Hyperbolic Graph Convolutional Network
• Quaternion Graph Neural Networks
• Mix Dimension in Poincaré Geometry for 3D Skeleton-based Action Recognition
• Switch Spaces: Learning Product Spaces with Sparse Gating
• Learning Hyperbolic Representations of Topological Features
• Unit Ball Model for Embedding Hierarchical Structures in the Complex Hyperbolic Space
• Hyperbolic Neural Networks++
• Are Hyperbolic Representations in Graphs Created Equal?
• Differentiating through the Fréchet Mean
• Lorentzian Graph Convolutional Networks
• Quaternion-Based Graph Convolution Network for Recommendation
• Linear Classifiers in Product Space Forms
• Graphs for deep learning representations
• HyperExpan: Taxonomy Expansion with Hyperbolic Representation Learning
• Isometric Graph Neural Networks