Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Robust subgaussian estimation with VC-dimension
(Submitted on 24 Apr 2020 (v1), last revised 8 Jul 2020 (this version, v3))
Abstract: Median-of-means (MOM) based procedures provide non-asymptotic and strong deviation bounds even when data are heavy-tailed and/or corrupted. This work proposes a new general way to bound the excess risk for MOM estimators. The core technique is the use of VC-dimension (instead of Rademacher complexity) to measure the statistical complexity. In particular, this allows to give the first robust estimators for sparse estimation which achieves the so-called subgaussian rate only assuming a finite second moment for the uncorrupted data. By comparison, previous works using Rademacher complexities required a number of finite moments that grows logarithmically with the dimension. With this technique, we derive new robust sugaussian bounds for mean estimation in any norm. We also derive a new robust estimator for covariance estimation that is the first to achieve subgaussian bounds without $L_4-L_2$ norm equivalence.
Submission history
From: Jules Depersin [view email][v1] Fri, 24 Apr 2020 13:21:09 GMT (40kb)
[v2] Fri, 5 Jun 2020 08:59:18 GMT (41kb)
[v3] Wed, 8 Jul 2020 16:16:13 GMT (41kb)
Link back to: arXiv, form interface, contact.