Structured Sequence Modeling with Graph Convolutional Recurrent Networks

TL;DR

The paper proposes Graph Convolutional Recurrent Networks (GCRN), a novel deep learning model that extends RNNs to graph-structured data, improving sequence prediction in applications like video frames and natural language processing.

Contribution

It introduces GCRN, combining CNNs on graphs with RNNs to effectively model structured sequences, a significant advancement over traditional RNNs.

Findings

GCRN improves prediction accuracy on structured sequence tasks.

GCRN accelerates learning speed compared to standard RNNs.

Exploiting spatial and dynamic data enhances model performance.

Abstract

This paper introduces Graph Convolutional Recurrent Network (GCRN), a deep learning model able to predict structured sequences of data. Precisely, GCRN is a generalization of classical recurrent neural networks (RNN) to data structured by an arbitrary graph. Such structured sequences can represent series of frames in videos, spatio-temporal measurements on a network of sensors, or random walks on a vocabulary graph for natural language modeling. The proposed model combines convolutional neural networks (CNN) on graphs to identify spatial structures and RNN to find dynamic patterns. We study two possible architectures of GCRN, and apply the models to two practical problems: predicting moving MNIST data, and modeling natural language with the Penn Treebank dataset. Experiments show that exploiting simultaneously graph spatial and dynamic information about data can improve both precision and learning speed.