Guardado en:
Detalles Bibliográficos
Autores principales: Diakonikolas, Ilias, Gao, Chao, Kane, Daniel M., Lafferty, John, Pensia, Ankit
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2510.10665
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912644183621632
author Diakonikolas, Ilias
Gao, Chao
Kane, Daniel M.
Lafferty, John
Pensia, Ankit
author_facet Diakonikolas, Ilias
Gao, Chao
Kane, Daniel M.
Lafferty, John
Pensia, Ankit
contents We study the task of noiseless linear regression under Gaussian covariates in the presence of additive oblivious contamination. Specifically, we are given i.i.d.\ samples from a distribution $(x, y)$ on $\mathbb{R}^d \times \mathbb{R}$ with $x \sim \mathcal{N}(0,\mathbf{I}_d)$ and $y = x^\top β+ z$, where $z$ is drawn independently of $x$ from an unknown distribution $E$. Moreover, $z$ satisfies $\mathbb{P}_E[z = 0] = α>0$. The goal is to accurately recover the regressor $β$ to small $\ell_2$-error. Ignoring computational considerations, this problem is known to be solvable using $O(d/α)$ samples. On the other hand, the best known polynomial-time algorithms require $Ω(d/α^2)$ samples. Here we provide formal evidence that the quadratic dependence in $1/α$ is inherent for efficient algorithms. Specifically, we show that any efficient Statistical Query algorithm for this task requires VSTAT complexity at least $\tildeΩ(d^{1/2}/α^2)$.
format Preprint
id arxiv_https___arxiv_org_abs_2510_10665
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Information-Computation Tradeoffs for Noiseless Linear Regression with Oblivious Contamination
Diakonikolas, Ilias
Gao, Chao
Kane, Daniel M.
Lafferty, John
Pensia, Ankit
Data Structures and Algorithms
Statistics Theory
Machine Learning
We study the task of noiseless linear regression under Gaussian covariates in the presence of additive oblivious contamination. Specifically, we are given i.i.d.\ samples from a distribution $(x, y)$ on $\mathbb{R}^d \times \mathbb{R}$ with $x \sim \mathcal{N}(0,\mathbf{I}_d)$ and $y = x^\top β+ z$, where $z$ is drawn independently of $x$ from an unknown distribution $E$. Moreover, $z$ satisfies $\mathbb{P}_E[z = 0] = α>0$. The goal is to accurately recover the regressor $β$ to small $\ell_2$-error. Ignoring computational considerations, this problem is known to be solvable using $O(d/α)$ samples. On the other hand, the best known polynomial-time algorithms require $Ω(d/α^2)$ samples. Here we provide formal evidence that the quadratic dependence in $1/α$ is inherent for efficient algorithms. Specifically, we show that any efficient Statistical Query algorithm for this task requires VSTAT complexity at least $\tildeΩ(d^{1/2}/α^2)$.
title Information-Computation Tradeoffs for Noiseless Linear Regression with Oblivious Contamination
topic Data Structures and Algorithms
Statistics Theory
Machine Learning
url https://arxiv.org/abs/2510.10665