bibkey: lidar2025dadqc authors: D. A. Lidar year: 2025 title: “Digital-Analog-Digital Quantum Supremacy” doi: 10.48550/arXiv.2512.07127 url: https://arxiv.org/abs/2512.07127v1 claim: “Let be a family of hardware graphs as in \cref{def:QPU-R-E}. For each , draw and . For any fixed choice of single-qubit angles in \cref{eq:HZ’}, there exist constants (depending only on ) such that for every , ” strata_touched:
- D5/S3/Quantum/Measurements/IQP/LidarRestrictedAnticoncentrationRefutation license: citation-only triage: anchor
Digital-Analog-Digital Quantum Supremacy
D. A. Lidar’s paper studies QPU-restricted IQP circuits and states an anticoncentration conjecture for every fixed hardware-graph family and fixed one-qubit angles.
Verified locator
DOI: https://doi.org/10.48550/arXiv.2512.07127
Source: https://arxiv.org/abs/2512.07127v1
Definition 1 (p. 4), verbatim from the source TeX:
Let
\mH_n=(\mV_{\mH},\mE_{\mH})be a fixed simple -regular hardware graph on labeled vertices, and let be a fixed constant (-independent). The \emph{-factors} of a graph\mHare \mF_d(\mH) \equiv {\mG=(\mV,\mE_{\mG}) : \mE_{\mG}\subseteq \mE_{\mH},\deg_{\mG}(v)=d\ \forall v\in\mV}. The \emph{QPU-restricted graph ensemble} is the probability space\mathrm{Unif}[\mF_d(\mH_n)]in which, for each run of\UDAD, a graph\mG\in\mF_d(\mH_n)is drawn uniformly at random.
Eqs. (17)–(20) specify the factor set, random-angle layer, circuit with its Hamiltonian, and anticoncentration event. Conjecture 2 (p. 4, labelled conj:host-anticonc) reads:
Let
\{\mH_n\}be a family of hardware graphs as in \cref{def:QPU-R-E}. For each , draw\mG\sim \mathrm{Unif}[\mF_d(\mH_n)]and . For any fixed choice of single-qubit angles in \cref{eq:HZ’}, there exist constants (depending only on ) such that for every ,\Pr_{\mG,\theta} \bigl[ P_{\UIQP^{(\theta)}}(s)\ge a 2^{-n} \bigr] \ge b.
The source’s Theorem 2 assumes Conjecture 2. The Lean settlement uses the fixed choice , the cyclic twin-pair hardware family with , and proves the negation of the conjecture’s conclusion.