arXiv:2609.39802
Double Localization for Quantum Gibbs Sampler Gaps: From an Abstract Framework to Finite-Group Models
Ryu Hayakawa, Angus Southwell, Caesnan M. G. Leditto, Kuo-Chin Chen, Min-Hsiu Hsieh
Abstract
We introduce double localization, a framework for proving spectral gaps of quantum Gibbs samplers through two complementary operations: geometric localization, which selects updates supported in small spatial regions, and interface localization, which focuses on a model-defined subspace of observables while remaining geometrically global. Our abstract gap theorem combines a global bound on this subspace, called the interface, with local quantum gap estimates under quantitative coupling and geometric assembly conditions. The resulting explicit lower bound controls the gap of the full dynamics without requiring the interface to be invariant under the generator. This separation allows global estimates, including classical comparison, to be combined with local control of the remaining quantum directions. We apply the framework to Hamiltonians built from the vertex terms of Kitaev's finite-group quantum double construction, without plaquette terms, on finite simple three-regular graphs of girth at least six. For the specified local Davies dynamics, we prove an unconditional spectral-gap lower bound by a positive constant independent of graph size for every fixed nontrivial finite group and each fixed inverse temperature $β$ with $0\leβJ<\log(5/3)$, where $J>0$ is the coupling strength. For the smallest non-Abelian group $S_3$, double localization yields a positive lower bound uniform in both graph size and the entire interval $0\leβJ\le\log 3$, providing a guarantee that does not follow directly from existing results. This demonstrates how model-specific finite calculations can extend spectral-gap guarantees for the full quantum dynamics.
AI Summary
The paper asks whether spatially local estimates and a global estimate on a model-selected observable subspace can jointly bound a quantum Gibbs sampler’s spectral gap, without requiring that subspace to be invariant under the dynamics. Its “double localization” framework combines local update collections with an interface projection $\Pi$. Given an interface gap $a$, local gaps, quantitative fixed-space coverage and overlap conditions, and local Schur-complement bounds controlling coupling to the complement, it proves $\operatorname{gap}(K)\ge(1-\sqrt{1-\eta})\min\{a,\tau\}$, where $\tau$ and $\eta$ are assembled from those inputs. The application treats $H=-J\sum_v A_v$, containing only the vertex projectors of finite-group quantum double models, on finite simple cubic graphs of girth at least six. For a specified Davies bath using edgewise left/right translations and group-basis projectors, the full quantum gap is bounded below independently of graph size for every fixed nontrivial finite group when $0\le\beta J<\log(5/3)$. For $S_3$, the guarantee is uniform throughout $0\le\beta J\le\log3$. The proof compares the irrep-label interface with a classical heat-bath chain, then controls complementary quantum directions using three-edge paths, local kernel analysis, and geometric assembly. The $S_3$ extension uses exact conditioned-star certificates and path coupling beyond the certified one-site Dobrushin range. The application excludes plaquette terms and depends on the specified bath and graph geometry; constants need not be uniform in group size. The abstract theorem does not require commutativity, but its verification here does. No modified logarithmic Sobolev inequality is established, and the temperature endpoints are certification limits, not demonstrated gap-closing thresholds.
were used to assist with research development, manuscript preparation, and review. The Python code for the finite verification was generated by GPT-5.6 Sol and verified by the authors.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 76 pages