CayleyNets: Graph Convolutional Neural Networks With Complex Rational Spectral Filters (2017-05-22T00:00:00.000000Z)

TL;DR

A new spectral domain convolutional architecture for deep learning on graphs with a new class of parametric rational complex functions (Cayley polynomials) allowing to efficiently compute spectral filters on graphs that specialize on frequency bands of interest.

Abstract

The rise of graph-structured data such as social networks, regulatory networks, citation graphs, and functional brain networks, in combination with resounding success of deep learning in various applications, has brought the interest in generalizing deep learning models to non-Euclidean domains. In this paper, we introduce a new spectral domain convolutional architecture for deep learning on graphs. The core ingredient of our model is a new class of parametric rational complex functions (Cayley polynomials) allowing to efficiently compute spectral filters on graphs that specialize on frequency bands of interest. Our model generates rich spectral filters that are localized in space, scales linearly with the size of the input data for sparsely connected graphs, and can handle different constructions of Laplacian operators. Extensive experimental results show the superior performance of our approach, in comparison to other spectral domain convolutional architectures, on spectral image classification, community detection, vertex classification, and matrix completion tasks.

Authors

M. Bronstein

14 papers

Federico Monti

5 papers

X. Bresson

7 papers

TL;DR

Abstract

Authors

References43 items

Fourier analysis

Filter Design for Autoregressive Moving Average Graph Filters

Graph Attention Networks

A Generalization of Convolutional Neural Networks to Graph-Structured Data

Geometric Matrix Completion with Recurrent Multi-Graph Neural Networks

Distance Metric Learning Using Graph Convolutional Networks: Application to Functional Brain Networks

Geometric Deep Learning on Graphs and Manifolds Using Mixture Model CNNs

Geometric Deep Learning: Going beyond Euclidean data

Semi-Supervised Classification with Graph Convolutional Networks

Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering

Learning Multiagent Communication with Backpropagation

Learning shape correspondence with anisotropic convolutional neural networks

Anisotropic Diffusion Descriptors

TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

Autoregressive Moving Average Graph Filtering

Collaborative Filtering with Graph Information: Consistency and Scalable Methods

Gated Graph Sequence Neural Networks

Diffusion-Convolutional Neural Networks

Convolutional Networks on Graphs for Learning Molecular Fingerprints

Learning class‐specific descriptors for deformable shapes using localized spectral convolutional networks

Deep Convolutional Networks on Graph-Structured Data

Geodesic Convolutional Neural Networks on Riemannian Manifolds

Adam: A Method for Stochastic Optimization

Matrix Completion on Graphs

Spectral Networks and Locally Connected Networks on Graphs

Speedup Matrix Completion with Side Information: Application to Multi-Label Learning

Discrete Signal Processing on Graphs: Frequency Analysis

Provable Inductive Matrix Completion

The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains

Discrete Signal Processing on Graphs

Selecting Receptive Fields in Deep Networks

Learning Fast Approximations of Sparse Coding

Collective Classification in Network Data

Exact Matrix Completion via Convex Optimization

Weighted Graph Cuts without Eigenvectors A Multilevel Approach

A new model for learning in graph domains

MovieLens unplugged: experiences with an occasionally connected recommender system

Economic inequality and burden-sharing in the provision of local environmental quality

Matrix Analysis, 2nd Ed

The Graph Neural Network Model

Collective Classi!cation in Network Data

Gradient-based learning applied to document recognition

Noname manuscript No. (will be inserted by the editor) Error Forgetting of Bregman Iteration

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names