Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering. Defferrard, M., Bresson, X., & Vandergheynst, P. arXiv:1606.09375 [cs, stat], February, 2017. arXiv: 1606.09375![Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex In this work, we are interested in generalizing convolutional neural networks (CNNs) from low-dimensional regular grids, where image, video and speech are represented, to high-dimensional irregular domains, such as social networks, brain connectomes or words' embedding, represented by graphs. We present a formulation of CNNs in the context of spectral graph theory, which provides the necessary mathematical background and efficient numerical schemes to design fast localized convolutional filters on graphs. Importantly, the proposed technique offers the same linear computational complexity and constant learning complexity as classical CNNs, while being universal to any graph structure. Experiments on MNIST and 20NEWS demonstrate the ability of this novel deep learning system to learn local, stationary, and compositional features on graphs.
@article{defferrard_convolutional_2017,
title = {Convolutional {Neural} {Networks} on {Graphs} with {Fast} {Localized} {Spectral} {Filtering}},
url = {http://arxiv.org/abs/1606.09375},
abstract = {In this work, we are interested in generalizing convolutional neural networks (CNNs) from low-dimensional regular grids, where image, video and speech are represented, to high-dimensional irregular domains, such as social networks, brain connectomes or words' embedding, represented by graphs. We present a formulation of CNNs in the context of spectral graph theory, which provides the necessary mathematical background and efficient numerical schemes to design fast localized convolutional filters on graphs. Importantly, the proposed technique offers the same linear computational complexity and constant learning complexity as classical CNNs, while being universal to any graph structure. Experiments on MNIST and 20NEWS demonstrate the ability of this novel deep learning system to learn local, stationary, and compositional features on graphs.},
urldate = {2022-03-02},
journal = {arXiv:1606.09375 [cs, stat]},
author = {Defferrard, Michaël and Bresson, Xavier and Vandergheynst, Pierre},
month = feb,
year = {2017},
note = {arXiv: 1606.09375},
keywords = {Computer Science - Machine Learning, Statistics - Machine Learning},
}
Downloads: 0
{"_id":"d2t4r7Y56Wr5ByJQ9","bibbaseid":"defferrard-bresson-vandergheynst-convolutionalneuralnetworksongraphswithfastlocalizedspectralfiltering-2017","author_short":["Defferrard, M.","Bresson, X.","Vandergheynst, P."],"bibdata":{"bibtype":"article","type":"article","title":"Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering","url":"http://arxiv.org/abs/1606.09375","abstract":"In this work, we are interested in generalizing convolutional neural networks (CNNs) from low-dimensional regular grids, where image, video and speech are represented, to high-dimensional irregular domains, such as social networks, brain connectomes or words' embedding, represented by graphs. We present a formulation of CNNs in the context of spectral graph theory, which provides the necessary mathematical background and efficient numerical schemes to design fast localized convolutional filters on graphs. Importantly, the proposed technique offers the same linear computational complexity and constant learning complexity as classical CNNs, while being universal to any graph structure. Experiments on MNIST and 20NEWS demonstrate the ability of this novel deep learning system to learn local, stationary, and compositional features on graphs.","urldate":"2022-03-02","journal":"arXiv:1606.09375 [cs, stat]","author":[{"propositions":[],"lastnames":["Defferrard"],"firstnames":["Michaël"],"suffixes":[]},{"propositions":[],"lastnames":["Bresson"],"firstnames":["Xavier"],"suffixes":[]},{"propositions":[],"lastnames":["Vandergheynst"],"firstnames":["Pierre"],"suffixes":[]}],"month":"February","year":"2017","note":"arXiv: 1606.09375","keywords":"Computer Science - Machine Learning, Statistics - Machine Learning","bibtex":"@article{defferrard_convolutional_2017,\n\ttitle = {Convolutional {Neural} {Networks} on {Graphs} with {Fast} {Localized} {Spectral} {Filtering}},\n\turl = {http://arxiv.org/abs/1606.09375},\n\tabstract = {In this work, we are interested in generalizing convolutional neural networks (CNNs) from low-dimensional regular grids, where image, video and speech are represented, to high-dimensional irregular domains, such as social networks, brain connectomes or words' embedding, represented by graphs. We present a formulation of CNNs in the context of spectral graph theory, which provides the necessary mathematical background and efficient numerical schemes to design fast localized convolutional filters on graphs. Importantly, the proposed technique offers the same linear computational complexity and constant learning complexity as classical CNNs, while being universal to any graph structure. Experiments on MNIST and 20NEWS demonstrate the ability of this novel deep learning system to learn local, stationary, and compositional features on graphs.},\n\turldate = {2022-03-02},\n\tjournal = {arXiv:1606.09375 [cs, stat]},\n\tauthor = {Defferrard, Michaël and Bresson, Xavier and Vandergheynst, Pierre},\n\tmonth = feb,\n\tyear = {2017},\n\tnote = {arXiv: 1606.09375},\n\tkeywords = {Computer Science - Machine Learning, Statistics - Machine Learning},\n}\n\n","author_short":["Defferrard, M.","Bresson, X.","Vandergheynst, P."],"key":"defferrard_convolutional_2017","id":"defferrard_convolutional_2017","bibbaseid":"defferrard-bresson-vandergheynst-convolutionalneuralnetworksongraphswithfastlocalizedspectralfiltering-2017","role":"author","urls":{"Paper":"http://arxiv.org/abs/1606.09375"},"keyword":["Computer Science - Machine Learning","Statistics - Machine Learning"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/mxmplx","dataSources":["aXmRAq63YsH7a3ufx"],"keywords":["computer science - machine learning","statistics - machine learning"],"search_terms":["convolutional","neural","networks","graphs","fast","localized","spectral","filtering","defferrard","bresson","vandergheynst"],"title":"Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering","year":2017}