Graph Neural Networks with Pytorch

Overview

This repository contains the following things.

• key explanations for the GNN algorithms introduced in papers
• brief descriptions of how to use the related pytorch code

More contents and applications will be added soon.

If you want more details, check the pytorch geometric github link.

---

Posts

Basic

Explanation of Message Passing base class

Explanation of Graph Fourier Transform

Paper Review and Code of Metapath2vec: Scalable Representation Learning for Heterogeneous Networks (KDD 2017)

GNN

Code of GCN: Semi-Supervised Classification with Graph Convolutional Networks (ICLR 2017)

Code and Paper Review of GraphSAGE: Inductive Representation Learning on Large Graphs (NIPS 2017)

Paper Review of PinSAGE: Graph Convolutional Neural Networks for Web-Scale Recommender Systems (2018)

Code and Paper Review of GAT: Graph Attention Networks (ICLR 2018)

Paper Review of SRGNN: Session-based Recommendation with Graph Neural Networks (2018)

Code of RGCN: Modeling Relational Data with Graph Convolutional Networks (ESWC 2018)

Code of SG: Simplifying Graph Convolutional Networks (CoRR 2019)

Code and Paper Review of GIN: Graph Isomorphism Networks (ICLR 2019)

Paper Review of GTN: Graph Transformer Networks (NIPS 2020)

Paper Review of HGT: Heterogenous Graph Transformer (WWW 2020)

Paper Review of Graphormer: Do Transformers Really Perform Bad for Graph Representation? (NIPS 2021)

Matrix Completion

Paper Review of GCMC: Graph Convolutional Matrix Completion (2017)

Paper Review of IGMC: Inductive Matrix Completion Based on Graph Neural Networks (ICLR 2020)

Pre-training

Paper Review of PCRec: Pre-training Graph Neural Network for Cross Domain Recommendation (2021)

Paper Review of Strategies for pre-training Grapn Neural Networks (ICLR 2020)

Pooling/Operator/Utilities

Paper Review of gPool: Graph U-Nets (ICML 2019) and DiffPool: Hierarchical Graph Representation Learning with Differentiable Pooling (NIPS 2018) and EigenPooling: Graph Convolutional Networks with EigenPooling (KDD 2019)

Code and Paper Review of APPNP: Predict then propagate, Graph Neural Networks meet Personalized PageRank (ICLR 2019)

Code and Paper Review of GDC: Diffusion Improves Graph Learning (NIPS 2019)

Scalability

Code and Paper Review of ClusterGCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks (KDD 2019)

Paper Review and code of SIGN: Scalable Inception Graph Neural Networks (CoRR 2020)

---

Applications

Exploring Reddit data