Structured Sequence Modeling with Graph Convolutional Recurrent Networks

Youngjoo Seo, Michaël Defferrard, Pierre Vandergheynst, Xavier Bresson
12/22/2016
stat.MLcs.LG

Official code: youngjoo-epfl/gconvRNN (181 stars, MIT, last commit 2018-03-15).

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.

View on arXivView PDF

Code Implementations(1)

youngjoo-epfl/gconvRNNOfficial100%
MIT

Graph convolutional recurrent neural network

18140PythonAug 28, 20178 years agoMIT

Similar Papers

An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling

84% match1 repo4.6k2018

Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks

83% match1 repo8922015

A Generative Self-Supervised Framework using Functional Connectivity in fMRI Data

83% match2023

EvolveGCN: Evolving Graph Convolutional Networks for Dynamic Graphs

83% match2 repos6202019

LLaGA: Large Language and Graph Assistant

83% match2 repos1642024

Multi-Graph Convolution Network for Pose Forecasting

83% match2023

Cite this paper

@article{seo2016structured,
  title  = {Structured Sequence Modeling with Graph Convolutional Recurrent Networks},
  author = {Youngjoo Seo and Michaël Defferrard and Pierre Vandergheynst and Xavier Bresson},
  year   = {2016},
  eprint = {1612.07659},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1612.07659}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1612.07659.svg)](https://codeofpaper.com/paper/1612.07659)

Discussion