English

Learning Entangled Single-Sample Distributions via Iterative Trimming

Machine Learning 2020-07-08 v2 Machine Learning

Abstract

In the setting of entangled single-sample distributions, the goal is to estimate some common parameter shared by a family of distributions, given one \emph{single} sample from each distribution. We study mean estimation and linear regression under general conditions, and analyze a simple and computationally efficient method based on iteratively trimming samples and re-estimating the parameter on the trimmed sample set. We show that the method in logarithmic iterations outputs an estimation whose error only depends on the noise level of the αn\lceil \alpha n \rceil-th noisiest data point where α\alpha is a constant and nn is the sample size. This means it can tolerate a constant fraction of high-noise points. These are the first such results for the method under our general conditions. It also justifies the wide application and empirical success of iterative trimming in practice. Our theoretical results are complemented by experiments on synthetic data.

Keywords

Cite

@article{arxiv.2004.09563,
  title  = {Learning Entangled Single-Sample Distributions via Iterative Trimming},
  author = {Hui Yuan and Yingyu Liang},
  journal= {arXiv preprint arXiv:2004.09563},
  year   = {2020}
}

Comments

Updated on AISTAT 2020 camera-ready: added comments on existing work at the end of Section 3

R2 v1 2026-06-23T14:58:43.939Z