Batch List-Decodable Linear Regression via Higher Moments
Abstract
We study the task of list-decodable linear regression using batches. A batch is called clean if it consists of i.i.d. samples from an unknown linear regression distribution. For a parameter , an unknown -fraction of the batches are clean and no assumptions are made on the remaining ones. The goal is to output a small list of vectors at least one of which is close to the true regressor vector in -norm. [DJKS23] gave an efficient algorithm, under natural distributional assumptions, with the following guarantee. Assuming that the batch size satisfies and the number of batches is , their algorithm runs in polynomial time and outputs a list of vectors at least one of which is close to the target regressor. Here we design a new polynomial time algorithm with significantly stronger guarantees under the assumption that the low-degree moments of the covariates distribution are Sum-of-Squares (SoS) certifiably bounded. Specifically, for any constant , as long as the batch size is and the degree- moments of the covariates are SoS certifiably bounded, our algorithm uses batches, runs in polynomial-time, and outputs an -sized list of vectors one of which is close to the target. That is, our algorithm achieves substantially smaller minimum batch size and final error, while achieving the optimal list size. Our approach uses higher-order moment information by carefully combining the SoS paradigm interleaved with an iterative method and a novel list pruning procedure. In the process, we give an SoS proof of the Marcinkiewicz-Zygmund inequality that may be of broader applicability.
Cite
@article{arxiv.2503.09802,
title = {Batch List-Decodable Linear Regression via Higher Moments},
author = {Ilias Diakonikolas and Daniel M. Kane and Sushrut Karmalkar and Sihan Liu and Thanasis Pittas},
journal= {arXiv preprint arXiv:2503.09802},
year = {2025}
}