Notice
This is not the latest version of this item. The latest version can be found at:https://dspace.mit.edu/handle/1721.1/138223.5
The overlap gap property in principal submatrix recovery
dc.contributor.author | Gamarnik, David | |
dc.date.accessioned | 2021-11-29T16:27:32Z | |
dc.date.available | 2021-11-29T15:19:58Z | |
dc.date.available | 2021-11-29T16:27:32Z | |
dc.date.issued | 2021-09-27 | |
dc.identifier.issn | 0178-8051 | |
dc.identifier.uri | https://hdl.handle.net/1721.1/138223.2 | |
dc.description.abstract | Abstract We study support recovery for a $$k \times k$$ k × k principal submatrix with elevated mean $$\lambda /N$$ λ / N , hidden in an $$N\times N$$ N × N symmetric mean zero Gaussian matrix. Here $$\lambda >0$$ λ > 0 is a universal constant, and we assume $$k = N \rho $$ k = N ρ for some constant $$\rho \in (0,1)$$ ρ ∈ ( 0 , 1 ) . We establish that there exists a constant $$C>0$$ C > 0 such that the MLE recovers a constant proportion of the hidden submatrix if $$\lambda {\ge C} \sqrt{\frac{1}{\rho } \log \frac{1}{\rho }}$$ λ ≥ C 1 ρ log 1 ρ , while such recovery is information theoretically impossible if $$\lambda = o( \sqrt{\frac{1}{\rho } \log \frac{1}{\rho }} )$$ λ = o ( 1 ρ log 1 ρ ) . The MLE is computationally intractable in general, and in fact, for $$\rho >0$$ ρ > 0 sufficiently small, this problem is conjectured to exhibit a statistical-computational gap. To provide rigorous evidence for this, we study the likelihood landscape for this problem, and establish that for some $$\varepsilon >0$$ ε > 0 and $$\sqrt{\frac{1}{\rho } \log \frac{1}{\rho } } \ll \lambda \ll \frac{1}{\rho ^{1/2 + \varepsilon }}$$ 1 ρ log 1 ρ ≪ λ ≪ 1 ρ 1 / 2 + ε , the problem exhibits a variant of the Overlap-Gap-Property (OGP). As a direct consequence, we establish that a family of local MCMC based algorithms do not achieve optimal recovery. Finally, we establish that for $$\lambda > 1/\rho $$ λ > 1 / ρ , a simple spectral method recovers a constant proportion of the hidden submatrix. | en_US |
dc.publisher | Springer Berlin Heidelberg | en_US |
dc.relation.isversionof | https://doi.org/10.1007/s00440-021-01089-7 | en_US |
dc.rights | Creative Commons Attribution-Noncommercial-Share Alike | en_US |
dc.rights.uri | http://creativecommons.org/licenses/by-nc-sa/4.0/ | en_US |
dc.source | Springer Berlin Heidelberg | en_US |
dc.title | The overlap gap property in principal submatrix recovery | en_US |
dc.type | Article | en_US |
dc.identifier.citation | Gamarnik, David, Jagannath, Aukosh and Sen, Subhabrata. 2021. "The overlap gap property in principal submatrix recovery." | en_US |
dc.relation.journal | Probability theory and related fields | en_US |
dc.eprint.version | Author's final manuscript | en_US |
dc.type.uri | http://purl.org/eprint/type/JournalArticle | en_US |
eprint.status | http://purl.org/eprint/status/PeerReviewed | en_US |
dc.date.updated | 2021-11-25T05:00:55Z | |
dc.language.rfc3066 | en | |
dc.rights.holder | The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature | |
dspace.embargo.terms | Y | |
dspace.date.submission | 2021-11-25T05:00:54Z | |
mit.journal.volume | 181 | en_US |
mit.license | OPEN_ACCESS_POLICY | |
mit.metadata.status | Ready for final review | en_US |