Belief propagation on the random k-SAT model

Amin Coja-Oghlan, Noela Müller, Jean B. Ravelomanana

Research output: Contribution to journalArticleAcademicpeer-review

2 Citations (Scopus)
23 Downloads (Pure)

Abstract

Corroborating a prediction from statistical physics, we prove that the belief propagation message passing algorithm approximates the partition function of the random k-SAT model well for all clause/variable densities and all inverse temperatures for which a modest absence of long-range correlations condition is satisfied. This condition is known as “replica symmetry” in physics language. From this result we deduce that a replica symmetry breaking phase transition occurs in the random k-SAT model at low temperature for clause/variable densities below but close to the satisfiability threshold.

Original languageEnglish
Pages (from-to)3718-3796
Number of pages79
JournalAnnals of Applied Probability
Volume32
Issue number5
DOIs
Publication statusPublished - Oct 2022

Funding

Funding. The first author was supported by DFG CO 646/3 and DFG CO 646/4. The second author was supported in part by ERC-Grant 772606-PTRCSP. The third author was supported by DFG CO 646/4.

FundersFunder number
International Advisory Board of ERC Advanced Grant Project THUNDERR772606-PTRCSP
European Union's Horizon 2020 - Research and Innovation Framework Programme772606
Deutsche ForschungsgemeinschaftCO 646/4, CO 646/3

    Keywords

    • belief propagation
    • k-SAT
    • phase transition
    • replica symmetry breaking

    Fingerprint

    Dive into the research topics of 'Belief propagation on the random k-SAT model'. Together they form a unique fingerprint.

    Cite this