Computer Science > Machine Learning
[Submitted on 12 Aug 2016 (this version), latest version 16 Nov 2016 (v2)]
Title:Beyond Spectral: Tight Bounds for Planted Gaussians
View PDFAbstract:Given an $n\times n$ matrix with i.i.d. N(0,1) entries except for a hidden n^{0.5-\delta} x n^{0.5-\delta} submatrix where the entries are N(0,\sigma_1^2) ("planted Gaussian"), we give a polynomial-time algorithm to recover the hidden part for some \delta >0, provided \sigma_1^2>2. We also show a matching lower bound: there is no polynomial time Statistical Query algorithm for detecting a hidden part when \sigma_1^2\in (0,2], unless \delta=0. We present two algorithms for the upper bound, with different behaviors close to the threshold. The lower bound as well as the stronger upper bound depend on the chi-squared distance of the two distributions. We give extensions to the setting when the hidden part has entries from N(\mu,\sigma_1^2).
Submission history
From: Santosh Vempala [view email][v1] Fri, 12 Aug 2016 00:36:42 UTC (16 KB)
[v2] Wed, 16 Nov 2016 16:31:27 UTC (19 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.