GCN-based GAE

The repository implements GCN-based GAE for monitoring technology convergence. The example.py demonstrates how to use GCN to generate patent and technological keywords vectors and technology intensity coordinates.

With pip

This repository is built on the Package dgl version 0.4.3 and Pytorch with a GTX 1080 Ti graphics card.

DGL with cuda (e.g., cuda version 10.1) can be installed using pip as follows:

pip install dgl-cu101==0.4.3

DGL (CPU only) can be installed using pip as follows:

pip install dgl==0.4.3

Usage

If you want to customize the model to your own datasets, please go first to check the Sample folder for data formats.