Rational Neural Networks for Approximating Jump Discontinuities of Graph Convolution Operator

For node level graph encoding, a recent important state-of-art method is the graph convolutional networks (GCN), which nicely integrate local vertex features and graph topology in the spectral domain. However, current studies suffer from several drawbacks: (1) graph CNNs relies on Chebyshev polynomial approximation which results in oscillatory approximation at jump discontinuities; (2) Increasing the order of Chebyshev polynomial can reduce the oscillations issue, but also incurs unaffordable computational cost; (3) Chebyshev polynomials require degree $\Omega$(poly(1/$\epsilon$)) to approximate a jump signal such as $|x|$, while rational function only needs $\mathcal{O}$(poly log(1/$\epsilon$))\cite{liang2016deep,telgarsky2017neural}. However, it's non-trivial to apply rational approximation without increasing computational complexity due to the denominator. In this paper, the superiority of rational approximation is exploited for graph signal recovering. RatioanlNet is proposed to integrate rational function and neural networks. We show that rational function of eigenvalues can be rewritten as a function of graph Laplacian, which can avoid multiplication by the eigenvector matrix. Focusing on the analysis of approximation on graph convolution operation, a graph signal regression task is formulated. Under graph signal regression task, its time complexity can be significantly reduced by graph Fourier transform. To overcome the local minimum problem of neural networks model, a relaxed Remez algorithm is utilized to initialize the weight parameters. Convergence rate of RatioanlNet and polynomial based methods on jump signal is analyzed for a theoretical guarantee. The extensive experimental results demonstrated that our approach could effectively characterize the jump discontinuities, outperforming competing methods by a substantial margin on both synthetic and real-world graphs.

[1]  D. Newman Rational approximation to | x , 1964 .

[2]  S. Stigler Gergonne's 1815 paper on the design and analysis of polynomial regression experiments , 1974 .

[3]  D. Wulbert The Rational Approximation of Real Functions , 1978 .

[4]  H. Hochstadt Complex Analysis: An Introduction to the Theory of Analytic Functions of One Complex Variable; 3rd ed. (Lars V. Ahlfors) , 1980 .

[5]  M. Powell,et al.  Approximation theory and methods , 1984 .

[6]  R. Varga,et al.  On the bernstein conjecture in approximation theory , 1985 .

[7]  J. E. Glynn,et al.  Numerical Recipes: The Art of Scientific Computing , 1989 .

[8]  J. Boyd Chebyshev and Fourier Spectral Methods , 1989 .

[9]  V. Sunder,et al.  The Laplacian spectrum of a graph , 1990 .

[10]  David J. C. MacKay,et al.  Bayesian Interpolation , 1992, Neural Computation.

[11]  Stéphane Mallat,et al.  Matching pursuits with time-frequency dictionaries , 1993, IEEE Trans. Signal Process..

[12]  Gabriel Taubin,et al.  A signal processing approach to fair surface design , 1995, SIGGRAPH.

[13]  G. Lorentz,et al.  Constructive approximation : advanced problems , 1996 .

[14]  R. Tibshirani Regression Shrinkage and Selection via the Lasso , 1996 .

[15]  Y Iida,et al.  Transportation Network Analysis , 1997 .

[16]  George Eastman House,et al.  Sparse Bayesian Learning and the Relevan e Ve tor Ma hine , 2001 .

[17]  M. Newman Spread of epidemic disease on networks. , 2002, Physical review. E, Statistical, nonlinear, and soft matter physics.

[18]  Keh-Moh Lin Book Review: Numerical Methods for Scientists and Engineers. By H. M. Antia , 2002 .

[19]  L. Hood,et al.  A Genomic Regulatory Network for Development , 2002, Science.

[20]  Koby Crammer,et al.  Online Passive-Aggressive Algorithms , 2003, J. Mach. Learn. Res..

[21]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[22]  B. Ripley,et al.  Robust Statistics , 2018, Encyclopedia of Mathematical Geosciences.

[23]  A. Ng Feature selection, L1 vs. L2 regularization, and rotational invariance , 2004, Twenty-first international conference on Machine learning - ICML '04.

[24]  H. Zou,et al.  Regularization and variable selection via the elastic net , 2005 .

[25]  Jason Weston,et al.  A unified architecture for natural language processing: deep neural networks with multitask learning , 2008, ICML '08.

[26]  Pierre Vandergheynst,et al.  Wavelets on Graphs via Spectral Graph Theory , 2009, ArXiv.

[27]  A. Pentland,et al.  Life in the network: The coming age of computational social science: Science , 2009 .

[28]  R. Pachón Algorithms for polynomial and rational approximation , 2010 .

[29]  H. Cohen Numerical Approximation Methods , 2011 .

[30]  Michael G. Rabbat,et al.  Approximating signals supported on graphs , 2012, 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[31]  Vivien Marx,et al.  High-throughput anatomy: Charting the brain's networks , 2012, Nature.

[32]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[33]  Pascal Frossard,et al.  The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains , 2012, IEEE Signal Processing Magazine.

[34]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[35]  Pierre Vandergheynst,et al.  Vertex-Frequency Analysis on Graphs , 2013, ArXiv.

[36]  Joan Bruna,et al.  Spectral Networks and Locally Connected Networks on Graphs , 2013, ICLR.

[37]  Zhiyuan Liu,et al.  Learning Entity and Relation Embeddings for Knowledge Graph Completion , 2015, AAAI.

[38]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[39]  Joan Bruna,et al.  Deep Convolutional Networks on Graph-Structured Data , 2015, ArXiv.

[40]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[41]  Xavier Bresson,et al.  Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering , 2016, NIPS.

[42]  Pierre Vandergheynst,et al.  Geometric Deep Learning: Going beyond Euclidean data , 2016, IEEE Signal Process. Mag..

[43]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[44]  Matus Telgarsky,et al.  Neural Networks and Rational Functions , 2017, ICML.

[45]  R. Srikant,et al.  Why Deep Neural Networks for Function Approximation? , 2016, ICLR.

[46]  Geoff Boeing,et al.  OSMnx: New Methods for Acquiring, Constructing, Analyzing, and Visualizing Complex Street Networks , 2016, Comput. Environ. Urban Syst..

[47]  Xiao-Ming Wu,et al.  Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning , 2018, AAAI.

[48]  U. Feige,et al.  Spectral Graph Theory , 2015 .