Salvato in:
Dettagli Bibliografici
Autori principali: Pourali, Herbod, Hashemian, Sajjad, Ardeshir-Larijani, Ebrahim
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2511.22524
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912734473355264
author Pourali, Herbod
Hashemian, Sajjad
Ardeshir-Larijani, Ebrahim
author_facet Pourali, Herbod
Hashemian, Sajjad
Ardeshir-Larijani, Ebrahim
contents We introduce an expander-sketching framework for list-decodable linear regression that achieves sample complexity $\tilde{O}((d+\log(1/δ))/α)$, list size $O(1/α)$, and near input-sparsity running time $\tilde{O}(\mathrm{nnz}(X)+d^{3}/α)$ under standard sub-Gaussian assumptions. Our method uses lossless expanders to synthesize lightly contaminated batches, enabling robust aggregation and a short spectral filtering stage that matches the best known efficient guarantees while avoiding SoS machinery and explicit batch structure.
format Preprint
id arxiv_https___arxiv_org_abs_2511_22524
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle List-Decodable Regression via Expander Sketching
Pourali, Herbod
Hashemian, Sajjad
Ardeshir-Larijani, Ebrahim
Machine Learning
Discrete Mathematics
G.3; G.2
We introduce an expander-sketching framework for list-decodable linear regression that achieves sample complexity $\tilde{O}((d+\log(1/δ))/α)$, list size $O(1/α)$, and near input-sparsity running time $\tilde{O}(\mathrm{nnz}(X)+d^{3}/α)$ under standard sub-Gaussian assumptions. Our method uses lossless expanders to synthesize lightly contaminated batches, enabling robust aggregation and a short spectral filtering stage that matches the best known efficient guarantees while avoiding SoS machinery and explicit batch structure.
title List-Decodable Regression via Expander Sketching
topic Machine Learning
Discrete Mathematics
G.3; G.2
url https://arxiv.org/abs/2511.22524