ERM and RERM are optimal estimators for regression problems when malicious outliers corrupt the labels

DOI10.1214/20-EJS1754zbMath1453.62484arXiv1910.10923MaRDI QIDQ2209821

Geoffrey Chinot

Publication date: 5 November 2020

Published in: Electronic Journal of Statistics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1910.10923

zbMATH Keywords

outliersrobustnessminimax-rate-optimalityregularized empirical risk minimizers (RERM)

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08) Nonparametric robustness (62G35) Minimax procedures in statistical decision theory (62C20)

Related Items (3)

All-in-one robust estimator of the Gaussian mean ⋮ ERM and RERM are optimal estimators for regression problems when malicious outliers corrupt the labels ⋮ Unnamed Item

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Geometric median and robust estimation in Banach spaces
• Estimation and testing under sparsity. École d'Été de Probabilités de Saint-Flour XLV -- 2015
• A general decision theory for Huber's \(\epsilon\)-contamination model
• On the prediction performance of the Lasso
• Probability in Banach spaces. Isoperimetry and processes
• Oracle inequalities in empirical risk minimization and sparse recovery problems. École d'Été de Probabilités de Saint-Flour XXXVIII-2008.
• Regularization in kernel learning
• Robust m-estimators of multivariate location and scatter
• Asymptotic behavior of M-estimators for the linear model
• Smooth discrimination analysis
• Sub-Gaussian estimators of the mean of a random vector
• Regularization and the small-ball method. I: Sparse recovery
• Robust covariance and scatter matrix estimation under Huber's contamination model
• Optimal aggregation of classifiers in statistical learning.
• On the conditions used to prove oracle results for the Lasso
• Slope meets Lasso: improved oracle bounds and optimality
• Robust statistical learning with Lipschitz and convex loss functions
• Robust machine learning by median-of-means: theory and practice
• Mean estimation with sub-Gaussian rates in polynomial time
• ERM and RERM are optimal estimators for regression problems when malicious outliers corrupt the labels
• Algorithms of robust stochastic optimization based on mirror descent method
• Estimation bounds and sharp oracle inequalities of regularized procedures with Lipschitz loss functions
• Localized Gaussian width of \(M\)-convex hulls with applications to Lasso and convex aggregation
• Local Rademacher complexities and oracle inequalities in risk minimization. (2004 IMS Medallion Lecture). (With discussions and rejoinder)
• Optimal rates for the regularized least-squares algorithm
• Empirical minimization
• Local Rademacher complexities
• Learning theory estimates via integral operators and their approximations
• The Future of Data Analysis
• Support Vector Machines
• The Influence Curve and Its Role in Robust Estimation
• Robust Estimators in High-Dimensions Without the Computational Intractability
• High-Dimensional Probability
• 10.1162/1532443041424337
• Efficient Algorithms and Lower Bounds for Robust Linear Regression
• High-Dimensional Robust Mean Estimation in Nearly-Linear Time
• Upper and Lower Bounds for Stochastic Processes
• A General Qualitative Definition of Robustness

This page was built for publication: ERM and RERM are optimal estimators for regression problems when malicious outliers corrupt the labels