A Bernstein-type inequality for functions of bounded interaction
classification
🧮 math.PR
keywords
functionsinequalityapplicationsargumentargumentsbernsteinbernstein-typebounded
read the original abstract
We give a distribution-dependent concentration inequality for functions of independent variables. The result extends Bernstein's inequality from sums to more general functions, whose variation in any argument does not depend too much on the other arguments. Applications sharpen existing bounds for U-statistics and the generalization error of regularized least squares.
This paper has not been read by Pith yet.
Forward citations
Cited by 1 Pith paper
-
Group-invariant Coresets for Data-efficient Active Learning
GRINCO performs acquisition in the quotient space induced by a transformation group using invariant embeddings or canonical representatives, pairs it with orbit-averaged loss, derives a generalization bound, and repor...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.