Graph connection Laplacian methods can be made robust to noise

DOI10.1214/14-AOS1275MaRDI QIDQ5963525zbMATH OpenOpenAlexFDO

Authors Noureddine El Karoui, H.-T. Wu

Publication date 22 February 2016

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1405.6231, https://projecteuclid.org/euclid.aos/1452004789

zbMATH Keywords

noisekernel methodsrobustnessrandom matricesspectral geometryconcentration of measurevector diffusion mapsgraph connection Laplacian methods

Mathematics Subject Classification ID

Random matrices (probabilistic aspects) (60B20) Limit theorems in probability theory (60F99)

Abstract: Recently, several data analytic techniques based on connection graph laplacian (CGL) ideas have appeared in the literature. At this point, the properties of these methods are starting to be understood in the setting where the data is observed without noise. We study the impact of additive noise on these methods, and show that they are remarkably robust. As a by-product of our analysis, we propose modifications of the standard algorithms that increase their robustness to noise. We illustrate our results in numerical simulations.

Recommendations

• Graph connection Laplacian and random matrices with random blocks
• Perturbation of the eigenvectors of the graph Laplacian: application to image denoising
• Weighted graph Laplace operator under topological noise
• Learning Theory
• Towards a theoretical foundation for Laplacian-based manifold methods

Cites work

• scientific article; zbMATH DE number 5542504 (Why is no real title available?)
• scientific article; zbMATH DE number 47363 (Why is no real title available?)
• scientific article; zbMATH DE number 967931 (Why is no real title available?)
• A Local Clustering Algorithm for Connection Graphs
• A class of generalized Laplacians on vector bundles devoted to multi-channel image processing
• A perfect match condition for point-set matching problems using the optimal mass transport approach
• A theoretical and computational framework for isometry invariant recognition of point cloud data
• Alternating projection, ptychographic imaging and phase synchronization
• Angular synchronization by eigenvectors and semidefinite programming
• Asymptotic Statistics
• Asymptotics of graphical projection pursuit
• Comparing Top k Lists
• Concentration of measure and spectra of random matrices: applications to correlation matrices, elliptical distributions and beyond
• Continuous Procrustes distance between two surfaces
• Eigenvector synchronization, graph rigidity and the molecule problem
• Geometric Representation of High Dimension, Low Sample Size Data
• Introduction to the mathematics of medical imaging
• On information plus noise kernel random matrices
• On robust regression with high-dimensional predictors
• Operator norm consistent estimation of large-dimensional sparse covariance matrices
• Phase retrieval with polarization
• Posterior convergence rates for Dirichlet mixtures of beta densities
• Ranking and sparsifying a connection graph
• Representation theoretic patterns in three dimensional cryo-electron microscopy. I: The intrinsic reconstitution algorithm
• The concentration of measure phenomenon
• Two-dimensional tomography from noisy projections taken at unknown random directions
• Vector diffusion maps and the connection Laplacian
• Viewing angle classification of cryo-electron microscopy images using eigenvectors

Cited in

(18)

• The geometry of synchronization problems and learning group actions
• Eigen-convergence of Gaussian kernelized graph Laplacian by manifold heat interpolation
• Doubly stochastic normalization of the Gaussian kernel is robust to heteroskedastic noise
• Spatiotemporal analysis using Riemannian composition of diffusion operators
• Latent common manifold learning with alternating diffusion: analysis and applications
• On the beneficial effect of noise in vertex localization
• Explore Intrinsic Geometry of Sleep Dynamics and Predict Sleep Stage by Unsupervised Learning Techniques
• The Steerable Graph Laplacian and its Application to Filtering Image Datasets
• Embeddings of Riemannian manifolds with finite eigenvector fields of connection Laplacian
• Think globally, fit locally under the manifold setup: asymptotic analysis of locally linear embedding
• Learning low-dimensional nonlinear structures from high-dimensional noisy data: an integral operator approach
• Solving jigsaw puzzles by the graph connection Laplacian
• The diffusion geometry of fibre bundles: horizontal diffusion maps
• Spectral convergence of the connection Laplacian from random samples
• Wave-shape oscillatory model for nonstationary periodic time series analysis
• Robust Inference of Manifold Density and Geometry by Doubly Stochastic Scaling
• Embedding Riemannian manifolds by the heat kernel of the connection Laplacian
• Approximations of the connection Laplacian spectra

This page was built for publication: Graph connection Laplacian methods can be made robust to noise

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5963525)