Doubly Stochastic Normalization of the Gaussian Kernel Is Robust to Heteroskedastic Noise

DOI10.1137/20M1342124zbMath1468.62431arXiv2006.00402OpenAlexW3137945549MaRDI QIDQ4999363

Boris Landa, Yuval Kluger, Ronald R. Coifman

Publication date: 6 July 2021

Published in: SIAM Journal on Mathematics of Data Science (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2006.00402

zbMATH Keywords

Gaussian kernelmanifold learningdoubly stochasticgraph Laplacianaffinity matrixheteroskedastic noise

Mathematics Subject Classification ID

Nonparametric robustness (62G35) Applications of statistics to biology and medical sciences; meta analysis (62P10) Protein sequences, DNA sequences (92D20) Statistical aspects of big data and data science (62R07)

Related Items

Eigen-convergence of Gaussian kernelized graph Laplacian by manifold heat interpolation, Scaling positive random matrices: concentration and asymptotic convergence, Robust Inference of Manifold Density and Geometry by Doubly Stochastic Scaling, A low discrepancy sequence on graphs, Spectral Convergence of Diffusion Maps: Improved Error Bounds and an Alternative Normalization

Uses Software

• t-SNE
• SpectralNet

Cites Work

• Unnamed Item
• Unnamed Item
• A new derivation of eigenvalue inequalities for the multinomial distribution
• On information plus noise kernel random matrices
• Wavelets on graphs via spectral graph theory
• Order independence and factor convergence in iterative scaling
• Variable bandwidth diffusion kernels
• The collected works of Wassily Hoeffding. Ed. by N. I. Fisher and P. K. Sen
• Clipped noisy images: heteroskedastic modeling and practical denoising
• Error estimates for spectral convergence of the graph Laplacian on random geometric graphs toward the Laplace-Beltrami operator
• Poisson noise reduction with non-local PCA
• Perturbation of the eigenvectors of the graph Laplacian: application to image denoising
• Role of normalization in spectral clustering for stochastic blockmodels
• Diffusion maps
• Diffusion wavelets
• Diffusion maps, spectral clustering and reaction coordinates of dynamical systems
• From graph to manifold Laplacian: the convergence rate
• Concerning nonnegative matrices and doubly stochastic matrices
• Diffusion Interpretation of Nonlocal Neighborhood Filters for Signal Denoising
• The Sinkhorn–Knopp Algorithm: Convergence and Applications
• Graph Laplacian Regularization for Image Denoising: Analysis in the Continuous Domain
• Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
• Manifold learning with bi-stochastic kernels
• The Steerable Graph Laplacian and its Application to Filtering Image Datasets
• Learning Theory
• A Review of Image Denoising Algorithms, with a New One
• On Stochastic Limit and Order Relationships
• Graph connection Laplacian methods can be made robust to noise