Debiasing the Lasso under Weaker Tail Assumptions

2026/08/05 by Leonardo Voltarelli, Roberto Imbuzeiro Oliveira
Mathematics · #math.ST #stat.TH #msc:62J07 #msc:62E20

paper · pdf

arxiv created 2026/08/05 · arxiv updated 2026/08/06

Abstract

We consider the problem of high-dimensional inference with the lasso estimator. Different methods including 'double selection' techniques and multiple versions of the 'debiased lasso' have been proposed for this task with noticeable success. However, most guarantees assume strong hypotheses on the underlying data process and the errors in the linear regression model, such as subgaussian designs and independence between errors and the data itself. We show that 'standardizing' one's dataset -- a natural procedure in practical penalized regression -- leads to same results under much weaker hypotheses, paying only a small price for not assuming light tails. The key technical point allowed by this step is exploiting the concentration properties of self-normalized processes. Importantly, we prove our results for two different methods closely related to the 'debiased lasso'. The second method performs valid inference even for a misspecified linear model, under mild sparsity conditions similar to the 'double selection' literature.

Citations

• Improving Genomic Prediction using High-dimensional Secondary Phenotypes: the Genetic Latent Factor Approach
• A remark on moment-dependent phase transitions in high-dimensional Gaussian approximations
• High-Dimensional Inference for Generalized Linear Models with Hidden Confounding
• De-biased Lasso for Generalized Linear Models with A Diverging Number of Covariates
• De-Biasing The Lasso With Degrees-of-Freedom Adjustment
• Robust high dimensional factor models with applications to statistical machine learning
• Debiasing the Debiased Lasso with Bootstrap
• De-biasing the Lasso: Optimal Sample Size for Gaussian Designs
• Confidence Intervals for High-Dimensional Linear Regression: Minimax Rates and Adaptivity
• A General Theory of Hypothesis Tests and Confidence Regions for Sparse High Dimensional Models
• A general theory of hypothesis tests and confidence regions for sparse high dimensional models
• Confidence intervals for high-dimensional inverse covariance estimation
• The lower tail of random quadratic forms, with applications to ordinary least squares and restricted eigenvalue properties
• High-Dimensional Methods and Inference on Structural and Treatment Effects
• Confidence Intervals and Hypothesis Testing for High-Dimensional Regression
• On asymptotically optimal confidence regions and tests for high-dimensional models
• Confidence Intervals for Low-Dimensional Parameters in High-Dimensional Linear Models
• Sparse Models and Methods for Optimal Instruments with an Application to Eminent Domain
• Restricted Eigenvalue Conditions on Subgaussian Random Matrices
• On the conditions used to prove oracle results for the Lasso
• Regularized Multivariate Regression for Identifying Master Predictors with Application to Integrative Genomics Study of Breast Cancer
• Simultaneous analysis of Lasso and Dantzig selector
• The Dantzig selector: Statistical estimation when p is much larger than n
• Near Optimal Signal Recovery From Random Projections: Universal Encoding Strategies?