Quantum process tomography of a compressed time evolution circuit on superconducting quantum processors

Maria Dincă, David J. Luitz, Maxime Debertolis · 2025

Read Quantum process tomography of a compressed time evolution circuit on superconducting quantum processors by Maria Dincă, David J. Luitz, Maxime Debertolis (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-2509-25342.txt.

Introduction

Maria Dinca, David Luitz, and Maxime Debertolis describe the noise of superconducting quantum processors with quantum process tomography. Quantum hardware today is held back by many kinds of noise. Any near-term quantum gain thus rests on methods built to resist noise.

The team used top tomography methods under real runtime limits on IBM quantum processors. Their main test compares Trotter time evolution for three-qubit and four-qubit wave functions. They set that next to a compressed circuit form of the same evolution. By looking at the spectral traits of the two channels, they find the compressed circuit gives larger eigenvalue moduli each time. That shows the compressed form resists noise better. The work gives users a way to check gains on real machines.

Catalog record

CREATORMaria Dincă, David J. Luitz, Maxime Debertolis
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-25342

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Quantum process tomography of a compressed time evolution circuit on superconducting quantum processors Maria Dincă, David J. Luitz, Maxime Debertolis Abstract As present day quantum hardware is limited by various noise mechanisms, quantum advantage can only be reached in the near-term by designing noise-resilient quantum algorithms. In this work, we employ state-of-the-art quantum process tomography (QPT) techniques to characterize the noise channels of IBM quantum processors under realistic runtime constraints. As our main application, we compare the Trotter time-evolution of three- and four-qubit wave functions to a compressed quantum circuit version of the same evolution operator. By analysing the spectral properties of the two process channels, we find that the compressed circuit systematically yields larger eigenvalue moduli, demonstrating better noise resilience. I Introduction Quantum processors are progressing at a remarkable pace and are widely anticipated to eventually surpass the computational power of classical computers Feynman 1982 ; Lloyd 1996 ; Shor 1997 ; Huang et al. 2022 ; King et al. 2024 . While the long-term goal of a fully error-corrected, fault-tolerant quantum computer remains out of reach with current technology, Noisy Intermediate-Scale Quantum (NISQ) devices provide an essential platform for exploring near-term quantum applications Preskill 2018 ; Boixo et al. 2018 ; Bharti et al. 2022 . These intermediate devices have proven over the years to be capable of performing meaningful experiments in many fields of physics Aspuru-Guzik et al. 2005 ; Lanyon et al. 2011 ; Heras et al. 2014 ; Salathé et al. 2015 ; Acín et al. 2018 ; Perdomo-Ortiz et al. 2018 ; Zhong et al. 2020 ; Quantum et al. 2020 ; Satzinger et al. 2021 ; L. et al. 2022 ; Richter and Pal 2021 , despite their inherent noise and limited qubit numbers. Among the variety of experimental platforms under active development, trapped ions Haffner et al. 2008 ; Ballance et al. 2016 , superconducting circuits Barends et al. 2014 ; Huang et al. 2020a and neutral atom platforms exploiting atomic Rydberg states Briegel et al. 2000 ; Bluvstein et al. 2022 have emerged as the most promising candidates. Trapped-ion qubits offer long coherence times and high-fidelity entangling operations, but are limited by relatively slow gate times and difficulties in scaling up the architecture. Neutral atom platforms face similar scaling challenges but bring the advantage of free reconfiguration of qubits to implement arbitrary connectivity Bluvstein et al. 2022 . Silicon spin qubits can be integrated with existing CMOS technology, allowing for scalability via reduced physical sizes, but achieving flexible connectivity is still a challenging task Siegel et al. 2025 . In contrast, superconducting platforms work with a fixed processor topology, but benefit from fast gate operations, and ongoing progress in scaling to larger qubit arrays. This makes them attractive for exploring algorithms and simulation protocols. One of the most promising near-term applications of quantum hardware is the simulation of quantum many-body systems Zalka 1998 ; Fauseweh 2024 , where the exponential complexity of classical computations is naturally avoided on quantum processors. However, current superconducting devices suffer from a number of limitations: qubit connectivity is typically restricted to nearest neighbors on some kind of two dimensional lattice, and every quantum gate introduces errors due to the required coupling with the environment. As more gates are applied, these errors accumulate and eventually dominate the computation, rendering the output indistinguishable from random noise. Several strategies have been developed to deal with this intrinsic vulnerability. A major direction is quantum error mitigation Temme et al. 2017 ; Cai et al. 2023 , which encompasses protocols to estimate or suppress noise without requiring the space-time overhead of full quantum error correction. These approaches include extrapolation techniques Li and Benjamin 2017 ; Cai 2021 or probabilistic error cancellation McDonough et al. 2022 ; Takagi et al. 2022 ; van den Berg et al. 2023 ; Gupta et al. 2024 , all of which aim to average out errors by performing additional measurements. Combining digital quantum simulation with error-mitigation techniques has already yielded encouraging results, allowing the reproduction of physically relevant processes such as the real-time evolution of interacting quantum systems Kim et al. 2023 ; Fauseweh 2024 . Although powerful, these methods come with significant measurement overheads. A complementary approach to reducing noise is to directly shorten the depth of quantum circuits, thereby limiting the number of noisy operations. This has motivated the development of compressed circuits Mansuroglu et al. 2023a ; Mansuroglu et al. 2023b ; Mc Keever and Lubasch 2023 ; Tepaske et al. 2023 ; Tepaske et al. 2024 , which provide optimized approximations of a given process—such as Hamiltonian time evolution—using parameterized gate sequences of tunable depth and connectivity. Circuit compression algorithms, typically designed and executed on classical computers, identify optimal representations that can reduce the depth by nearly 50% compared to standard Trotter–Suzuki decompositions at fixed fidelity Tepaske et al. 2023 . In principle, such reductions are especially valuable for superconducting platforms, where non-local interactions require additional SWAP gates, which are particularly noisy and can lead to significant error propagation. So far, compressed circuits have been mostly studied in theory or numerical simulations, under simplified noise models. Their practical performance under realistic hardware noise remains largely untested. In this article, we address this open question by confronting compressed circuits with standard Trotter decompositions in experimental realizations on quantum hardware. To quantitatively assess the quality of the implemented processes, we make use of quantum process tomography (QPT) Chuang and Nielsen 1997 ; O’Brien et al. 2004 ; Howard et al. 2006 ; Mohseni et al. 2008 ; Nielsen and Chuang 2010 . QPT reconstructs the quantum channel describing the implemented process, without relying on assumptions about input states or device-specific noise, thus providing an unbiased characterization of the experiment. From the reconstructed process matrix, one can not only compute fidelities with respect to target processes but also identify sources of decoherence or systematic errors in the implementation Kofman and Korotkov 2009 ; Kimmel et al. 2014 ; Flammia and Wallman 2020 ; van den Berg et al. 2023 ; Onorati et al. 2023 . However, QPT scales poorly with system size, as the number of parameters describing the channel grows exponentially with the number of qubits. For this reason, our study focuses on small systems, where QPT is still feasible, but already reveals clear differences between circuit implementations. Specifically, we demonstrate the use of compressed circuits on three- and four-qubit Heisenberg spin chains. For the three-qubit case, we perform a full process tomography to characterize the implemented processes in detail. For four qubits, where the full-QPT becomes computationally prohibitive, we employ Pauli twirling Wallman and Emerson 2016 ; Knill 2004 and selective quantum process tomography (SQPT) Bendersky et al. 2008 ; Bendersky et al. 2009 ; Schmiegelow et al. 2011 ; Perito et al. 2018 ; Pears Stefano et al. 2023 , which allows the efficient reconstruction of relevant elements of the process matrix. This combination enables us to benchmark compressed circuits against Trotter decompositions under realistic noise conditions and to demonstrate the advantage of circuit compression in experimental quantum simulations. To the best of our knowledge, this is the first experimental implementation of SQPT beyond three qubits and for deep-circuit channels rather than

Read the complete text