Robust machine learning by median-of-means: theory and practice
From MaRDI portal
Publication:2196199
DOI10.1214/19-AOS1828zbMath1487.62034arXiv1711.10306OpenAlexW3010140118MaRDI QIDQ2196199
Matthieu Lerasle, Guillaume Lecué
Publication date: 28 August 2020
Published in: The Annals of Statistics (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/1711.10306
Nonparametric regression and quantile regression (62G08) Estimation in multivariate analysis (62H12) Ridge regression; shrinkage estimators (Lasso) (62J07) Minimax procedures in statistical decision theory (62C20) Learning and adaptive systems in artificial intelligence (68T05)
Related Items
On least squares estimation under heteroscedastic and heavy-tailed errors, Robust sub-Gaussian estimation of a mean vector in nearly linear time, Robust Bregman clustering, The main contributions of robust statistics to statistical science and a new challenge, All-in-one robust estimator of the Gaussian mean, Super-polynomial accuracy of one dimensional randomized nets using the median of means, Optimal robust mean and location estimation via convex programs with respect to any pseudo-norms, Robust multivariate mean estimation: the optimality of trimmed mean, Robust statistical learning with Lipschitz and convex loss functions, Data perturbations in stochastic generalized equations: statistical robustness in static and sample average approximated models, Byzantine-robust distributed sparse learning for \(M\)-estimation, High-dimensional \(M\)-estimation for Byzantine-robust decentralized learning, A robust estimator of the proportional hazard transform for massive data, Robust partially linear trend filtering for regression estimation and structure discovery, Non-asymptotic analysis and inference for an outlyingness induced winsorized mean, Topics in robust statistical learning, Unnamed Item, Robust supervised learning with coordinate gradient descent, Robust classification via MOM minimization, ERM and RERM are optimal estimators for regression problems when malicious outliers corrupt the labels, Outlier detection in networks with missing links, A MOM-based ensemble method for robustness, subsampling and hyperparameter tuning, Iteratively reweighted \(\ell_1\)-penalized robust regression, Robust \(k\)-means clustering for distributions with two moments, K-bMOM: A robust Lloyd-type clustering algorithm based on bootstrap median-of-means, Scale calibration for high-dimensional robust regression, Mean estimation and regression under heavy-tailed distributions: A survey, Unnamed Item, Learning in Repeated Auctions, Distribution-free robust linear regression
Uses Software
Cites Work
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- On asymptotically optimal confidence regions and tests for high-dimensional models
- Geometric median and robust estimation in Banach spaces
- SLOPE is adaptive to unknown sparsity and asymptotically minimax
- Rho-estimators for shape restricted density estimation
- A new method for estimation and model selection: \(\rho\)-estimation
- Sub-Gaussian mean estimators
- Statistics for high-dimensional data. Methods, theory and applications.
- Oracle inequalities in empirical risk minimization and sparse recovery problems. École d'Été de Probabilités de Saint-Flour XXXVIII-2008.
- Segmentation of the mean of heteroscedastic data via cross-validation
- Estimator selection with respect to Hellinger-type risks
- Robust linear least squares regression
- Stabilité et instabilité du risque minimax pour des variables indépendantes équidistribuées
- Risk bounds for statistical learning
- SLOPE-adaptive variable selection via convex optimization
- A survey of cross-validation procedures for model selection
- Lasso-type recovery of sparse representations for high-dimensional data
- Random generation of combinatorial structures from a uniform distribution
- Asymptotic methods in statistical decision theory
- The space complexity of approximating the frequency moments
- On optimality of empirical risk minimization in linear aggregation
- Regularization and the small-ball method. I: Sparse recovery
- Challenging the empirical mean and empirical variance: a deviation study
- Robust low-rank matrix estimation
- Slope meets Lasso: improved oracle bounds and optimality
- Learning from MOM's principles: Le Cam's approach
- Distributed statistical estimation and rates of convergence in normal approximation
- Risk minimization by median-of-means tournaments
- Estimation bounds and sharp oracle inequalities of regularized procedures with Lipschitz loss functions
- Simultaneous analysis of Lasso and Dantzig selector
- Regularization, sparse recovery, and median-of-means tournaments
- Statistical performance of support vector machines
- High-dimensional generalized linear models and the lasso
- Sup-norm convergence rate and sign concentration property of Lasso and Dantzig estimators
- Confidence sets in sparse regression
- Model selection via testing: an alternative to (penalized) maximum likelihood estimators.
- High-dimensional graphs and variable selection with the Lasso
- Convergence of estimates under dimensionality restrictions
- Learning without Concentration
- Choice of V for V-Fold Cross-Validation in Least-Squares Density Estimation
- The Future of Data Analysis
- Bounding the Smallest Singular Value of a Random Matrix Without Concentration
- The Influence Curve and Its Role in Robust Estimation
- Regularization and the small-ball method II: complexity dependent error rates
- Estimation of High Dimensional Mean Regression in the Absence of Symmetry and Light Tail Assumptions
- Weakly decomposable regularization penalties and structured sparsity
- Robust Estimation of a Location Parameter
- A General Qualitative Definition of Robustness
- Confidence Intervals for Low Dimensional Parameters in High Dimensional Linear Models
- Robust Statistics
- A unified framework for high-dimensional analysis of \(M\)-estimators with decomposable regularizers
