Non-Haar random circuits form unitary designs as fast as Haar random circuits
Toshihiro Yada, Ryotaro Suzuki, Yosuke Mitsuhashi, Nobuyuki Yoshioka · 2025
Read Non-Haar random circuits form unitary designs as fast as Haar random circuits by Toshihiro Yada, Ryotaro Suzuki, Yosuke Mitsuhashi, Nobuyuki Yoshioka (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-07390.txt.
Introduction
Random quantum circuits scramble information so thoroughly that their behavior mimics a perfectly uniform random operation, a property called a unitary design. Toshihiro Yada and co-authors ask how the choice of random parts affects the time that takes, in a 2025 paper. Truly uniform random gates are an idealization; real circuits use simpler randomizers.
The team proves the simpler circuits are not slower in any meaningful sense. The depth they need to form a design is at most a constant multiple of the ideal circuit's depth, whatever the system size. The result holds for chains, grids, long-range layouts, and even patchwork shallow circuits. The authors draw lessons for randomized benchmarking, random circuit sampling, and the physics of quantum chaos. Their message to experimenters is freedom: practical randomness is already fast enough.
Catalog record
| CREATOR | Toshihiro Yada, Ryotaro Suzuki, Yosuke Mitsuhashi, Nobuyuki Yoshioka |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-07390
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Non-Haar random circuits form unitary designs as fast as Haar random circuits Toshihiro Yada, Ryotaro Suzuki, Yosuke Mitsuhashi, Nobuyuki Yoshioka Abstract The unitary design formation in random circuits has attracted considerable attention due to its wide range of practical applications and relevance to fundamental physics. While the formation rates in Haar random circuits have been extensively studied in previous works, it remains an open question how these rates are affected by the choice of local randomizers. In this work, we prove that the circuit depths required for general non-Haar random circuits to form unitary designs are upper bounded by those for the corresponding Haar random circuits, up to a constant factor independent of the system size. This result is derived in a broad range of circuit structures, including one- and higher-dimensional lattices, geometrically non-local configurations, and even extremely shallow circuits with patchwork architectures. We provide specific applications of these results in randomized benchmarking and random circuit sampling, and also discuss their implications for quantum many-body physics. Our work lays the foundation for flexible and robust randomness generation in real-world experiments, and offers new insights into chaotic dynamics in complex quantum systems. Introduction: Random quantum circuits have been intensively studied as a powerful tool for various quantum information processing tasks such as randomized benchmarking Helsen et al. 2022 and quantum tomography Elben et al. 2023 ; Huang et al. 2020 , and also have served as tractable models for the nonequilibrium dynamics of quantum many-body systems Fisher et al. 2023 ; Nahum et al. 2017 . A central focus in these investigations is the rate at which the global randomness is generated. While generating an exact Haar random ensemble in an N -qudit system requires a circuit depth of at least e^{\Omega(N)} , an ensemble that mimics Haar randomness up to the t -th moment, called a unitary t -design, is known to be created much more efficiently, with a depth of \mathrm{poly}N Harrow and Low 2009 ; Brown and Viola 2010 ; Brandão et al. 2016 ; Haferkamp 2022 ; Chen et al. 2024 ; Haferkamp and Hunter-Jones 2021 ; Hunter-Jones 2019 ; Harrow and Mehraban 2023 . Furthermore, the recent intensive investigations reveal that \mathrm{poly}N -depth formation of unitary t -design is ubiquitous in generic circuit architectures Mittal and Hunter-Jones 2023 ; Belkin et al. 2024 , and even O(\log N) -depth formation is achievable in certain special geometries Schuster et al. 2024 ; LaRacuente and Leditzky 2024 , which we call patchwork circuit in this Letter. However, these previous works have mainly focused on scenarios where each two-qudit unitary gate that constitutes the random circuit is sampled from the Haar measure, or at least where the unitary gates have a special structure, such as forming a unitary subgroup Webb 2015 ; Zhu 2017 . A prominent question then arises: To what extent is the rate of unitary design formation affected by the choice of local randomizers? This question is interesting not only from a fundamental perspective but also from an experimental standpoint, since in real experiments unitary gates are typically chosen non-Haar randomly, for example, from a discrete gate set Boixo et al. 2018 ; Arute et al. 2019 ; Morvan et al. 2024 ; Zhu et al. 2022 ; Gao et al. 2025 . While this question has been addressed in seminal works Oszmaniec et al. 2021 ; Oszmaniec et al. 2024 , their depth upper bounds are loose by \mathrm{poly}N -factors and are valid only for circuits with specific structures, excluding any fixed-architecture circuits such as brickwork circuits. Therefore, the critical open problem is to clarify whether unitary designs can be formed with the same circuit depth order even when the local randomizers are modified, across a wide range of circuit structures. Figure 1: Schematics of (a) 1D local circuit, (b) 1D brickwork circuit, and (c) 1D patchwork circuit. The patchwork circuit is a special geometry introduced in Ref. Schuster et al. 2024 to generate a unitary design with extremely shallow circuit depth. (d) The best known upper bounds for the circuit depths required to form \varepsilon -approximate unitary t -designs in N -qudit systems. The upper bound for a single-layer-connected circuit obtained in this work is tighter than those in Refs. Oszmaniec et al. 2021 ; Oszmaniec et al. 2024 by a \mathrm{poly}N factor. Furthermore, we also derive depth bounds for a multilayer-connected circuit and a patchwork circuit composed of them, which were not covered in those works. The asterisk in the patchwork row indicates that this bound is based on the relative error definition of an approximate t -design, while the other rows use Definition 1 . In this work, we comprehensively study various circuit structures and prove that the circuit depths required to form unitary t -designs in general non-Haar random circuits are upper bounded by those in the corresponding Haar-random circuits, up to a constant prefactor. This prefactor is determined by the randomness strength of the two-qudit unitary ensembles constituting the circuit, and is independent of the system size N . When these ensembles contain universal gate sets, it can be further shown that this prefactor exhibits at most \mathrm{polylog}t scaling in t , indicating the insensitivity of the t -design formation depth to the choice of local randomizers in both N and t . In deriving these results, we classify the circuit structures into three types and develop different proof methods for each class. Specifically, the three classes are: (\mathrm{i}) single-layer-connected circuit , where the two-qudit gates available in a single layer form a connected graph covering all sites, as shown in Fig. 1 (a); (\mathrm{ii}) multilayer-connected circuit , where multiple layers are required to connect all sites, as shown in Fig. 1 (b); and (\mathrm{iii}) patchwork circuit , where small patches of normal random circuits are glued together, as shown in Fig. 1 (c). The proof methods developed in this work significantly improve both the tightness and the comprehensiveness of the existing upper bounds, as summarized in Fig. 1 (d). These results have diverse applications and rich implications in various fields. In terms of practical applications, our work enhances the flexibility of global randomness generation in real experiments and ensures that the generation rate is largely unaffected by deformations of local gates, for example, those caused by coherent errors. We raise a few specific examples that could benefit from this enhanced flexibility, in randomized benchmarking and random circuit sampling. From a fundamental physics perspective, our work implies that various chaotic phenomena, such as information scrambling Roberts and Yoshida 2017 ; Nahum et al. 2018 ; von Keyserlingk et al. 2018 and complexity growth Oszmaniec et al. 2024 ; Brandão et al. 2021 ; Haferkamp et al. 2022 , universally emerge with the same circuit depth order, independent of the choice of local randomness. Thus, our work lays the foundation for randomness generation in quantum many-body systems and opens up numerous avenues for future applications. Preliminary: Before addressing the random circuit, we briefly review the basic tools for an approximate unitary t -design. We here consider a probability distribution \nu on the set of unitary operations on q -dimensional Hilbert space, denoted as \mathrm{U}(q) . The distribution \nu is said to