🤖 AI Summary
This work addresses the detection phase transition in the stochastic block model (SBM) with $q=3$ and $q=4$ communities, aiming to resolve whether a “computational-statistical gap” exists between information-theoretic limits and computational feasibility. Leveraging a novel fine-grained coupling between graphs and broadcast processes on trees—combined with Galton–Watson tree reconstruction theory, information-theoretic lower bounds, and precise critical-point quantification—we rigorously establish that, for sufficiently large average degree, the information-theoretic undetectability threshold coincides exactly with the Kesten–Stigum (KS) bound for both $q=3$ and $q=4$, thereby eliminating any computational-statistical gap. Notably, we show that KS tightness for $q=4$ depends critically on the average degree, whereas it fails for $q geq 5$. These results fully confirm the physics conjecture of Decelle et al. (2011) and provide the first rigorous phase-transition benchmark for high-dimensional SBMs and tree reconstruction.
📝 Abstract
In this paper, we rigorously establish the predictions in ground breaking work in statistical physics by Decelle, Krzakala, Moore, Zdeborová (2011) regarding the block model, in particular in the case of q=3 and q=4 communities. We prove that for q=3 and q=4 there is no computational-statistical gap if the average degree is above some constant by showing that it is information theoretically impossible to detect below the Kesten-Stigum bound. We proceed by showing that for the broadcast process on Galton-Watson trees, reconstruction is impossible for q=3 and q=4 if the average degree is sufficiently large. This improves on the result of Sly (2009), who proved similar results for d-regular trees when q=3. Our analysis of the critical case q=4 provides a detailed picture showing that the tightness of the Kesten-Stigum bound in the antiferromagnetic regime depends on the average degree of the tree. We also prove that for q≥ 5, the Kesten-Stigum bound is not sharp. Our results prove conjectures of Decelle, Krzakala, Moore, Zdeborová (2011), Moore (2017), Abbe and Sandon (2018) and Ricci-Tersenghi, Semerjian, and Zdeborová (2019). Our proofs are based on a new general coupling of the tree and graph processes and on a refined analysis of the broadcast process on the tree.
