Fast Mean Estimation with Sub-Gaussian Rates

TL;DR

This paper introduces a computationally efficient estimator for the mean of a random vector that achieves sub-Gaussian error bounds under minimal assumptions, improving upon previous methods in speed and simplicity.

Contribution

The authors present a new mean estimator with sub-Gaussian rates that is faster and simpler than existing polynomial-time estimators based on sum-of-squares hierarchy.

Findings

Achieves sub-Gaussian error bounds with finite mean and covariance assumptions.

Computational complexity is $O(n^4 + n^2 d)$, faster than previous methods.

Simpler analysis compared to prior polynomial-time estimators.

Abstract

We propose an estimator for the mean of a random vector in $\mathbb{R}^d$ that can be computed in time $O(n^4+n^2d)$ for $n$ i.i.d.~samples and that has error bounds matching the sub-Gaussian case. The only assumptions we make about the data distribution are that it has finite mean and covariance; in particular, we make no assumptions about higher-order moments. Like the polynomial time estimator introduced by Hopkins, 2018, which is based on the sum-of-squares hierarchy, our estimator achieves optimal statistical efficiency in this challenging setting, but it has a significantly faster runtime and a simpler analysis.