The distribution of the Lasso: uniform control over sparse balls and adaptive parameter tuning

DOI10.1214/20-AOS2038zbMATH Open1480.62145arXiv1811.01212OpenAlexW3202433451MaRDI QIDQ2054498FDOQ2054498

Authors: Yanyan Li 

Publication date: 3 December 2021

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

Abstract: The Lasso is a popular regression method for high-dimensional problems in which the number of parameters $het{a}_1,dots,het{a}_N$, is larger than the number $n$ of samples: $N>n$. A useful heuristics relates the statistical properties of the Lasso estimator to that of a simple soft-thresholding denoiser,in a denoising problem in which the parameters $(het{a}_i{)}_{ileN}$ are observed in Gaussian noise, with a carefully tuned variance. Earlier work confirmed this picture in the limit $n,Noinfty$, pointwise in the parameters $heta$, and in the value of the regularization parameter. Here, we consider a standard random design model and prove exponential concentration of its empirical distribution around the prediction provided by the Gaussian denoising model. Crucially, our results are uniform with respect to $heta$ belonging to $el{l}_q$ balls, $qin[0,1]$, and with respect to the regularization parameter. This allows to derive sharp results for the performances of various data-driven procedures to tune the regularization. Our proofs make use of Gaussian comparison inequalities, and in particular of a version of Gordon's minimax theorem developed by Thrampoulidis, Oymak, and Hassibi, which controls the optimum value of the Lasso optimization problem. Crucially, we prove a stability property of the minimizer in Wasserstein distance, that allows to characterize properties of the minimizer itself.

Full work available at URL: https://arxiv.org/abs/1811.01212

Recommendations

• On the distribution of the adaptive LASSO estimator
• The Lasso as an \(\ell _{1}\)-ball model selection procedure
• A practical scheme and fast algorithm to tune the Lasso with optimality guarantees
• On the sensitivity of the Lasso to the number of predictor variables
• On the distribution, model selection properties and uniqueness of the Lasso estimator in low and high dimensions

zbMATH Keywords

cross-validationlinear regressionsparsity

Mathematics Subject Classification ID

Linear regression; mixed models (62J05) Ridge regression; shrinkage estimators (Lasso) (62J07)

Cites Work

• Title not available (Why is that?)
• Statistics for high-dimensional data. Methods, theory and applications.
• Estimation of the mean of a multivariate normal distribution
• On the conditions used to prove oracle results for the Lasso
• Simultaneous analysis of Lasso and Dantzig selector
• The Dantzig selector: statistical estimation when \(p\) is much larger than \(n\). (With discussions and rejoinder).
• Confidence Intervals and Hypothesis Testing for High-Dimensional Regression
• A statistical mechanics approach to de-biasing and uncertainty estimation in Lasso for random measurements
• Confidence Intervals for Low Dimensional Parameters in High Dimensional Linear Models
• A unified framework for high-dimensional analysis of \(M\)-estimators with decomposable regularizers
• On asymptotically optimal confidence regions and tests for high-dimensional models
• Hypothesis Testing in High-Dimensional Regression Under the Gaussian Random Design Model: Asymptotic Theory
• Degrees of freedom in lasso problems
• Optimal Transport
• Minimax Rates of Estimation for High-Dimensional Linear Regression Over $\ell_q$-Balls
• Minimax risk over \(l_ p\)-balls for \(l_ q\)-error
• Decoding by Linear Programming
• Neighborliness of randomly projected simplices in high dimensions
• The Estimation of Prediction Error
• Some inequalities for Gaussian processes and applications
• High-dimensional centrally symmetric polytopes with neighborliness proportional to dimension
• Title not available (Why is that?)
• Living on the edge: phase transitions in convex programs with random data
• High dimensional robust M-estimation: asymptotic variance via approximate message passing
• The LASSO Risk for Gaussian Matrices
• The Noise-Sensitivity Phase Transition in Compressed Sensing
• The Dynamics of Message Passing on Dense Graphs, with Applications to Compressed Sensing
• Convex Recovery of a Structured Signal from Independent Random Linear Measurements
• Compressive Phase Retrieval via Generalized Approximate Message Passing
• Precise Error Analysis of Regularized <inline-formula> <tex-math notation="LaTeX">$M$ </tex-math> </inline-formula>-Estimators in High Dimensions
• On the impact of predictor geometry on the performance on high-dimensional ridge-regularized generalized robust regression estimators
• Mutual Information and Optimality of Approximate Message-Passing in Random Linear Estimation
• On cross-validated Lasso in high dimensions
• Accuracy assessment for high-dimensional linear regression
• Non-Negative Principal Component Analysis: Message Passing Algorithms and Sharp Asymptotics
• Consistent parameter estimation for Lasso and approximate message passing
• A modern maximum-likelihood theory for high-dimensional logistic regression
• State evolution for approximate message passing with non-separable functions
• Optimal errors and phase transitions in high-dimensional generalized linear models
• Optimization-Based AMP for Phase Retrieval: The Impact of Initialization and $\ell_{2}$ Regularization

Cited In (20)

• Asymptotic normality of robust \(M\)-estimators with convex penalty
• Noise covariance estimation in multi-task high-dimensional linear models
• A practical scheme and fast algorithm to tune the Lasso with optimality guarantees
• A Unifying Tutorial on Approximate Message Passing
• Optimal linear discriminators for the discrete choice model in growing dimensions
• Sharp global convergence guarantees for iterative nonconvex optimization with random data
• Aggregated hold out for sparse linear regression with a robust loss function
• Universality of regularized regression estimators in high dimensions
• Noisy linear inverse problems under convex constraints: exact risk asymptotics in high dimensions
• Precise statistical analysis of classification accuracies for adversarial training
• Local convexity of the TAP free energy and AMP convergence for \(\mathbb{Z}_2\)-synchronization
• Generic error bounds for the generalized Lasso with sub-exponential data
• The Lasso as an \(\ell _{1}\)-ball model selection procedure
• Debiasing convex regularized estimators and interval estimation in linear models
• Consistent parameter estimation for Lasso and approximate message passing
• Consistent parameter estimation for Lasso and approximate message passing
• Surprises in high-dimensional ridgeless least squares interpolation
• Sensitivity of \(\ell_1\) minimization to parameter choice
• The Lasso with general Gaussian designs with applications to hypothesis testing
• Fundamental barriers to high-dimensional regression with convex penalties

This page was built for publication: The distribution of the Lasso: uniform control over sparse balls and adaptive parameter tuning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2054498)