Graph Neural Networks in TensorFlow and Keras with Spektral

TL;DR

Spektral is an open-source Python library that simplifies building graph neural networks with TensorFlow and Keras, offering a wide range of methods and utilities for graph deep learning tasks.

Contribution

It provides a comprehensive, user-friendly toolkit for creating and prototyping graph neural networks, suitable for both beginners and experts.

Findings

Spektral supports node classification, graph classification, and graph regression tasks.

The library demonstrates competitive performance on benchmark datasets.

It enables rapid development and experimentation with graph neural network models.

Abstract

In this paper we present Spektral, an open-source Python library for building graph neural networks with TensorFlow and the Keras application programming interface. Spektral implements a large set of methods for deep learning on graphs, including message-passing and pooling operators, as well as utilities for processing graphs and loading popular benchmark datasets. The purpose of this library is to provide the essential building blocks for creating graph neural networks, focusing on the guiding principles of user-friendliness and quick prototyping on which Keras is based. Spektral is, therefore, suitable for absolute beginners and expert deep learning practitioners alike. In this work, we present an overview of Spektral's features and report the performance of the methods implemented by the library in scenarios of node classification, graph classification, and graph regression.