caiolmart/node-concordance

Using GNNs to predict paper subjects

arXiv Paper Classification

In this project I will try to predict the subject of a collection of papers from arXiv using Graph Neural Networks.

The dataset, ogbn-arxiv is from the Open Graph Benchmark (OGB). Downloading instructions can be found here.

Requirements

This project uses Pytorch and Pytorch Geometric.

For the installation with GPU enabled a simple pip install -r requirements.txt may not work.

Follow the instructions on:

• https://pytorch.org/get-started/locally/
• https://pytorch-geometric.readthedocs.io/en/latest/notes/installation.html