Fast Graph Representation Learning with PyTorch Geometric
By Matthias Fey, Jan Eric Lenssen. arXiv.org, 2019
5,499 citations recorded by the source corpus.
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.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Graph Learning, ML Systems & Hardware
View Paper: https://arxiv.org/abs/1903.02428
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Yoshua Bengio
- Mohit Bansal
- Percy Liang
- Jianfeng Gao
- Sergey Levine
- Quoc V. Le
- Ruslan Salakhutdinov
- Luke S. Zettlemoyer
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Trevor Darrell
- Xiaogang Wang
- Ming-Hsuan Yang
- Jitendra Malik
- Andrea Vedaldi
- Andrew Zisserman
In Graph Learning
- Jure Leskovec
- Stephan Günnemann
- William L. Hamilton
- Matthias Fey
- Jian Tang
- Marinka Zitnik
- Aleksandar Bojchevski
- Michael M. Bronstein
In ML Systems & Hardware
- Alex Smola
- Eran Yahav
- Uri Alon
- Michael M. Bronstein
- Bernard Ghanem
- Guohao Li
- Yuxiao Dong
- Shuiwang Ji
Citations in the corpus
- Strategies for Pre-training Graph Neural Networks (2019, methodology): “We use Pytorch (Paszke et al., 2017) and Pytorch Geometric (Fey & Lenssen, 2019) for all of our implementation.”
- Auto-GNN: Neural architecture search of graph neural networks (2019, methodology): “e function: As shown in [24], aggregate function is crucial to capture neighborhood structure for learning node representation. Herein GNN architecture is developed based on package Pytorch Geometric [29]. The package provides the following available aggregate functions: {SUMMATION,MEAN,MAXPOOLING}.”
- Provably Powerful Graph Networks (2019, neutral): “Each molecule is represented by an adjacency matrix, a distance matrix (between atoms), categorical data on the edges, and node features; the data was obtained from the pytorch-geometric library (Fey and Lenssen, 2019).”
- Just Jump: Dynamic Neighborhood Aggregation in Graph Neural Networks (2019, methodology): “The code with all its evaluation examples is integrated into the PyTorch Geometric1 library (Fey & Lenssen, 2019).”
- Multi-Scale Representation Learning on Proteins (2022, methodology): “, 2019) using the PyTorch Geometric library (Fey and Lenssen, 2019).”
- Simple and Deep Graph Convolutional Networks (2020, methodology): “We reuse the metrics already reported in (Fey & Lenssen, 2019) for GCN, GAT, and APPNP, and the best metrics reported in (Rong et al., 2020) for JKNet, JKNet(Drop) and Incep(Drop).”
- Multi-scale Attributed Node Embedding (2019, methodology): “Each model was optimized with the Adam optimizer [38] with the standard moving average parameters and the model implementations are sparsity aware modifications based on PyTorch Geometric [39].”
- Graph Contrastive Learning with Adaptive Augmentation (2020, methodology): “s existing state-of-the-art methods and even surpasses several supervised counterparts. A IMPLEMENTATION DETAILS A.1 Computing Infrastructures All models are implemented using PyTorch Geometric 1.6.1 [8], PyTorch 1.6.0 [29], and NetworkX 2.5. All datasets used throughout experiments are available in”
- Utilizing graph machine learning within drug discovery and development (2021, background): “where γ , □ and φ are the update, aggregation and message passing functions, respectively, and (t) indicates the layer index [93].”
- P HYSICS - INFORMED DEEP NEURAL NETWORK FOR RIGID - BODY PROTEIN DOCKING (2022, methodology): “For data processing and batching, we use PyTorch Geometric (Fey & Lenssen, 2019).”