Fast Graph Representation Learning with PyTorch Geometric

Abstract

We introduce PyTorch Geometric, a library for deep learning on irregularly structured input data such as graphs, point clouds and manifolds, built upon PyTorch. In addition to general graph data structures and processing methods, it contains a variety of recently published methods from the domains of relational learning and 3D data processing. PyTorch Geometric achieves high data throughput by leveraging sparse GPU acceleration, by providing dedicated CUDA kernels and by introducing efficient mini-batch handling for input examples of different size. In this work, we present the library in detail and perform a comprehensive comparative study of the implemented methods in homogeneous evaluation scenarios.

ICLR 2019 (RLGM Workshop)

References in corpus (3)

• Simplifying Graph Convolutional Networks
• Variational Graph Auto-Encoders
• Just Jump: Dynamic Neighborhood Aggregation in Graph Neural Networks

Cited by in corpus (127)

• Towards Deeper Graph Neural Networks
• Deep Graph Contrastive Representation Learning
• Self-Attention Graph Pooling
• TUDataset: A collection of benchmark datasets for learning with graphs
• DeeperGCN: All You Need to Train Deeper GCNs
• Composition-based Multi-Relational Graph Convolutional Networks
• Accelerating 3D Deep Learning with PyTorch3D
• Combining Label Propagation and Simple Models Out-performs Graph Neural Networks
• Edge Contraction Pooling for Graph Neural Networks
• GPT-GNN: Generative Pre-Training of Graph Neural Networks
• Structure-Feature based Graph Self-adaptive Pooling
• Deep Graph Matching Consensus
• Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures
• Geoopt: Riemannian Optimization in PyTorch
• EEG-GCNN: Augmenting Electroencephalogram-based Neurological Disease Diagnosis using a Domain-guided Graph Convolutional Neural Network
• ReCoG: A Deep Learning Framework with Heterogeneous Graph for Interaction-Aware Trajectory Prediction
• Spectral-based Graph Convolutional Network for Directed Graphs
• Hierarchical Inter-Message Passing for Learning on Molecular Graphs
• Detecting Code Clones with Graph Neural Networkand Flow-Augmented Abstract Syntax Tree
• Directed Acyclic Graph Neural Networks
• Inductive Representation Learning on Temporal Graphs
• PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation
• Heterogeneous Graph Transformer
• Towards Robust Graph Contrastive Learning
• Scalable Graph Neural Networks for Heterogeneous Graphs
• Search to aggregate neighborhood for graph neural network
• Just Jump: Dynamic Neighborhood Aggregation in Graph Neural Networks
• GraphCL: Contrastive Self-Supervised Learning of Graph Representations
• ForceNet: A Graph Neural Network for Large-Scale Quantum Calculations
• Boost then Convolve: Gradient Boosting Meets Graph Neural Networks
• Learned Low Precision Graph Neural Networks
• Message Passing Networks for Molecules with Tetrahedral Chirality
• Understanding Isomorphism Bias in Graph Data Sets
• A Meta-Learning Approach for Graph Representation Learning in Multi-Task Settings
• Belief Propagation Neural Networks
• Balancing Multi-level Interactions for Session-based Recommendation
• Graph Neural Network for Interpreting Task-fMRI Biomarkers
• Graph Coarsening with Neural Networks
• DRIFT: Deep Reinforcement Learning for Functional Software Testing
• Speaker attribution with voice profiles by graph-based semi-supervised learning
• Understanding Graph Neural Networks from Graph Signal Denoising Perspectives
• FeatGraph: A Flexible and Efficient Backend for Graph Neural Network Systems
• Understanding graph embedding methods and their applications
• On Graph Classification Networks, Datasets and Baselines
• Tri-graph Information Propagation for Polypharmacy Side Effect Prediction
• Discriminative structural graph classification
• Bi-Level Graph Neural Networks for Drug-Drug Interaction Prediction
• Graph Convolutional Networks for traffic anomaly
• Optimization of Graph Neural Networks with Natural Gradient Descent
• HeterSkinNet: A Heterogeneous Network for Skin Weights Prediction
• MGN-Net: a multi-view graph normalizer for integrating heterogeneous biological network populations
• Revisiting Graph Convolutional Network on Semi-Supervised Node Classification from an Optimization Perspective
• Rubik: A Hierarchical Architecture for Efficient Graph Learning
• GENN: Predicting Correlated Drug-drug Interactions with Graph Energy Neural Networks
• SceneGraphFusion: Incremental 3D Scene Graph Prediction from RGB-D Sequences
• Benchmarking Graph Neural Networks on Link Prediction
• Chickenpox Cases in Hungary: a Benchmark Dataset for Spatiotemporal Signal Processing with Graph Neural Networks
• Simplifying Architecture Search for Graph Neural Network
• Are Graph Neural Networks Miscalibrated?
• A Long Horizon Planning Framework for Manipulating Rigid Pointcloud Objects
• RelEx: A Model-Agnostic Relational Model Explainer
• Graph Representation learning for Audio & Music genre Classification
• Self-supervised Representation Learning for Evolutionary Neural Architecture Search
• Ivy: Templated Deep Learning for Inter-Framework Portability
• Net2: A Graph Attention Network Method Customized for Pre-Placement Net Length Estimation
• Utilising Graph Machine Learning within Drug Discovery and Development
• GE-SpMM: General-purpose Sparse Matrix-Matrix Multiplication on GPUs for Graph Neural Networks
• Refactoring Policy for Compositional Generalizability using Self-Supervised Object Proposals
• Semi-supervised Anomaly Detection on Attributed Graphs
• Multitask Learning On Graph Neural Networks Applied To Molecular Property Predictions
• Uncertainty-aware Attention Graph Neural Network for Defending Adversarial Attacks
• Scene-Graph Augmented Data-Driven Risk Assessment of Autonomous Vehicle Decisions
• Policy-GNN: Aggregation Optimization for Graph Neural Networks
• Contextualised Graph Attention for Improved Relation Extraction
• Exploiting Edge-Oriented Reasoning for 3D Point-based Scene Graph Analysis
• Task-Agnostic Morphology Evolution
• Node2Seq: Towards Trainable Convolutions in Graph Neural Networks
• A pipeline for fair comparison of graph neural networks in node classification tasks
• Traffic4cast 2020 -- Graph Ensemble Net and the Importance of Feature And Loss Function Design for Traffic Prediction
• Attention-Based Learning on Molecular Ensembles
• Nudge Attacks on Point-Cloud DNNs
• Self-supervised Graph Representation Learning via Bootstrapping
• CommPOOL: An Interpretable Graph Pooling Framework for Hierarchical Graph Representation Learning
• Self-supervised Auxiliary Learning for Graph Neural Networks via Meta-Learning
• Mesorasi: Architecture Support for Point Cloud Analytics via Delayed-Aggregation
• GraphGallery: A Platform for Fast Benchmarking and Easy Development of Graph Neural Networks Based Intelligent Software
• PyTorch-Direct: Enabling GPU Centric Data Access for Very Large Graph Neural Network Training with Irregular Accesses
• Graph Classification by Mixture of Diverse Experts
• Explainable Link Prediction for Privacy-Preserving Contact Tracing
• HyGCN: A GCN Accelerator with Hybrid Architecture
• Breaking the Expressive Bottlenecks of Graph Neural Networks
• Graph Neural Networks Including Sparse Interpretability
• Efficient Colon Cancer Grading with Graph Neural Networks
• Deep learning of material transport in complex neurite networks
• Graph Traversal with Tensor Functionals: A Meta-Algorithm for Scalable Learning
• User Preference-aware Fake News Detection
• ARENA: Asynchronous Reconfigurable Accelerator Ring to Enable Data-Centric Parallel Computing
• When to Talk: Chatbot Controls the Timing of Talking during Multi-turn Open-domain Dialogue Generation
• Are Hyperbolic Representations in Graphs Created Equal?
• SGQuant: Squeezing the Last Bit on Graph Neural Networks with Specialized Quantization
• Graph Information Vanishing Phenomenon inImplicit Graph Neural Networks
• Optimizing Memory Efficiency of Graph Neural Networks on Edge Computing Platforms
• Symmetry-adapted graph neural networks for constructing molecular dynamics force fields
• Geometric Graph Representations and Geometric Graph Convolutions for Deep Learning on Three-Dimensional (3D) Graphs
• Wireless 3D Point Cloud Delivery Using Deep Graph Neural Networks
• Analyzing the Performance of Graph Neural Networks with Pipe Parallelism
• TreeRNN: Topology-Preserving Deep GraphEmbedding and Learning
• Combinatorial Learning of Graph Edit Distance via Dynamic Embedding
• Autonomous Exploration Under Uncertainty via Deep Reinforcement Learning on Graphs
• Gini in a Bottleneck: Sparse Molecular Representations for Graph Convolutional Neural Networks
• Catastrophic Forgetting in Deep Graph Networks: an Introductory Benchmark for Graph Classification
• Instance Segmentation GNNs for One-Shot Conformal Tracking at the LHC
• Decentralized Statistical Inference with Unrolled Graph Neural Networks
• Walk2Map: Extracting Floor Plans from Indoor Walk Trajectories
• Speaker diarization with session-level speaker embedding refinement using graph neural networks
• Contextual Scene Augmentation and Synthesis via GSACNet
• Pooling in Graph Convolutional Neural Networks
• Label-Consistency based Graph Neural Networks for Semi-supervised Node Classification
• Approximate Knowledge Graph Query Answering: From Ranking to Binary Classification
• Node Attribute Completion in Knowledge Graphs with Multi-Relational Propagation
• Isometric Graph Neural Networks
• Characterizing and Understanding GCNs on GPU
• Graph Neural Network for Hamiltonian-Based Material Property Prediction
• Fast Graph Learning with Unique Optimal Solutions
• Scalable Adversarial Attack on Graph Neural Networks with Alternating Direction Method of Multipliers
• Improving the Expressive Power of Graph Neural Network with Tinhofer Algorithm
• Spatio-Temporal Sparsification for General Robust Graph Convolution Networks