Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2010.14577 · New Journal of Physics

Bilinear dynamic mode decomposition for quantum control

Andy Goldschmidt, Eurika Kaiser, J L DuBois, S L Brunton, J N Kutz

Published Feb 24, 2021• 36 citations
arXivHTMLPDF

Abstract

Abstract Data-driven methods for establishing quantum optimal control (QOC) using time-dependent control pulses tailored to specific quantum dynamical systems and desired control objectives are critical for many emerging quantum technologies. We develop a data-driven regression procedure, bilinear dynamic mode decomposition (biDMD), that leverages time-series measurements to establish quantum system identification for QOC. The biDMD optimization framework is a physics-informed regression that makes use of the known underlying Hamiltonian structure. Further, the biDMD can be modified to model both fast and slow sampling of control signals, the latter by way of stroboscopic sampling strategies. The biDMD method provides a flexible, interpretable, and adaptive regression framework for real-time, online implementation in quantum systems. Further, the method has strong theoretical connections to Koopman theory, which approximates nonlinear dynamics with linear operators. In comparison with many machine learning paradigms minimal data is needed to construct a biDMD model, and the model is easily updated as new data is collected. We demonstrate the efficacy and performance of the approach on a number of representative quantum systems, showing that it also matches experimental results.

DOI
10.1088/1367-2630/abe972

Related papers

Quantum machine learning optimization using Koopman operator technique

Subodh Nath Pushpak, Sarika Jain, Siddharth Kalra

Quantum machine learning (QML) is a nascent field showing great potential in addressing complex problems. QML algorithms aim to combine the qubit’s properties, like entanglement, interference, and superposition, to perform better than any classical computation in specific tasks. This paper explores a new way of employing the Koopman operator to demonstrate its application to quantum optimization in quantum machine learning. Koopman operator-based quantum optimization has applications in various dynamical systems, where it can optimally capture the full nonlinear behavior of a system [1]. This work investigates the Koopman operator approach for quantum optimization for quantum machine learning algorithms. To do so, we capitalize on recent breakthroughs in the field, such as data-driven methods for uncovering Koopman models and the application of the Koopman operator to investigate quantum machine learning optimal strategies. Specifically, we provide a strategy-driven approach with the full pipeline for encoding classical dynamical systems into quantum-accessible representation, thus employing beneficial properties of Koopman and transfer operators. This strategy enables us to represent the state space of classical systems in terms suitable for quantum computation. Here, we introduce a framework for using the Koopman operator to represent complex, high-dimensional classical dynamics on a quantum computer to enhance classical machine learning algorithms by transferring their capabilities onto quantum systems. Additionally, we provide a programmatical framework for using the Koopman operator in quantum machine learning frameworks on the IBM Qiskit framework.

Periodicals of Engineering and Natural Sciences (PEN)•Apr 29, 2025•0 citations

Predicting quantum ground-state energy by data-driven Koopman analysis of variational parameter nonlinear dynamics

Nobuyuki Okuma

In recent years, the application of machine learning to physics has been actively explored. In this paper, we study a method for estimating the ground-state energy of quantum Hamiltonians by applying data-driven Koopman analysis within the framework of variational wave functions. Koopman theory is a framework for analyzing the nonlinear dynamics of vectors, in which the dynamics are linearized by lifting the vectors to functions defined over the original vector space. We focus on the fact that the imaginary-time Schrödinger equation, when restricted to a variational wave function, is described by a nonlinear time evolution of the variational parameter vector. We collect sample points of this nonlinear dynamics at parameter configurations where the discrepancy between the true imaginary-time dynamics and the dynamics on the variational manifold is small, and perform data-driven continuous Koopman analysis. Within our formulation, the ground-state energy is reduced to the leading eigenvalue of a differential operator known as the Koopman generator. As a concrete example, we generate samples for the four-site transverse-field Ising model and estimate the ground-state energy using extended dynamic mode decomposition (EDMD). Furthermore, as an extension of this framework, we formulate the method for the case where the variational wave function is given by a uniform matrix product state on an infinite chain. By employing computational techniques developed within the framework of the time-dependent variational principle, all the quantities required for our analysis, including error estimation, can be computed efficiently in such systems. Since our approach provides predictions for the ground-state energy even when the true ground state lies outside the variational manifold, it is expected to complement conventional variational methods.

cond-mat.str-elCross-listedarXiv:2603.23887•Mar 26, 2026•0 citations

Quantum optimal control of superconducting qubits based on machine-learning characterization

Élie Genois, Noah Stevenson, Noah Goss +2 more

Implementing fast and high-fidelity quantum operations using open-loop quantum optimal control relies on having an accurate model of the quantum dynamics. Any deviations between this model and the complete dynamics of the device, such as the presence of spurious modes or pulse distortions, can degrade the performance of optimal controls in practice. Here, we propose an experimentally simple approach to realize optimal quantum controls tailored to the device parameters and environment while specifically characterizing this quantum system. Concretely, we use physics-inspired machine learning to infer an accurate model of the dynamics from experimentally available data and then optimize our experimental controls on this trained model. We show the power and feasibility of this approach by optimizing arbitrary single-qubit operations on a superconducting transmon qubit, using detailed numerical simulations. We demonstrate that this framework produces an accurate description of the device dynamics under arbitrary controls, together with the precise pulses achieving arbitrary single-qubit gates with a high fidelity of about 99.99%.

arXiv:2410.22603•Physical Review Applied•Aug 19, 2025•15 citations

Open-loop linear control of quadratic Hamiltonians with applications

Mattias Johnsson, Daniel Klaus Burgarth

The quantum harmonic oscillator is one of the most fundamental objects in physics. We consider the case where it is extended to an arbitrary number modes and includes all possible terms that are bilinear in the annihilation and creation operators, and we assume that we also have an arbitrary time-dependent drive term that is linear in those operators. Such a Hamiltonian is very general, covering a broad range of systems, including quantum optics, superconducting circuit QED, quantum error correcting codes, Bose-Einstein condensates, atomic wave packet transport beyond the adiabatic limit, and many others. We examine this situation from the point of view of quantum control, making use of optimal control theory to determine what can be accomplished, both when the controls are arbitrary and when they must minimize some cost function. In particular we develop a class of analytical pulses. We then apply our theory to a number of specific topical physical systems to illustrate its use and provide explicit control functions, including the case of the continuously driven conditional displacement gate.

arXiv:2304.11776•Physical Review A•Jan 24, 2024•4 citations

Dynamical Mode Decomposition for Infinite-Dimensional Open Quantum Systems; Liouvillian Spectral Analysis and Parameter Estimation

Yuzuru Kato, Hiroya Nakao

Dynamic mode decomposition (DMD) is a data-driven method for the estimation, prediction, and control of complex dynamical systems, which has gained much attention in the fields of nonlinear dynamics and fluid mechanics. A DMD method for quantum spin systems, described by a linear dynamical system of a finite-dimensional set of observables, has been proposed recently. In this study, we propose two DMD methods applicable to infinite-dimensional open quantum systems, which use time-series data obtained by quantum state tomography. First, we propose a kernel DMD method for a data-driven spectral analysis of the Liouville superoperator. Second, we propose a method for the parameter estimation of the Liouville superoperator, which incorporates prior knowledge of the model structure into DMD. The proposed methods can accurately reconstruct the system dynamics and show that DMD frameworks can be applicable to infinite-dimensional open quantum systems.

IEEE Conference on Decision and Control (CDC)•Dec 13, 2023•0 citations

Optimal Control of Unknown Collective Spin Systems via a Neural Network Surrogate

Chen Yaofeng, Li You

Abstract Quantum optimal control (QOC) relies on accurately modeling system dynamics and is often challenged by unknown or inaccessible interactions in real systems. Taking an unknown collective spin system as an example, this work introduces a machine-learning-based, data-driven scheme to overcome the challenges encountered, with a trained neural network (NN) assuming the role of a surrogate model that captures the system’s dynamics and subsequently enables QOC to be performed on the NN instead of on the real system. The trained NN surrogate proves effective for practical QOC tasks and is further demonstrated to be adaptable to different experimental conditions, remaining robust across varying system sizes and pulse durations.

Chinese Physics Letters•Aug 15, 2025•2 citations
Abstract Data-driven methods for establishing quantum optimal control (QOC) using time-dependent control pulses tailored to specific quantum dynamical systems and desired control objectives are critical for many emerging quantum technologies. We develop a data-driven regression procedure, bilinear dynamic mode decomposition (biDMD), that leverages time-series measurements to establish quantum system identification for QOC. The biDMD optimization framework is a physics-informed regression that makes use of the known underlying Hamiltonian structure. Further, the biDMD can be modified to model both fast and slow sampling of control signals, the latter by way of stroboscopic sampling strategies. The biDMD method provides a flexible, interpretable, and adaptive regression framework for real-time, online implementation in quantum systems. Further, the method has strong theoretical connections to Koopman theory, which approximates nonlinear dynamics with linear operators. In comparison with many machine learning paradigms minimal data is needed to construct a biDMD model, and the model is easily updated as new data is collected. We demonstrate the efficacy and performance of the approach on a number of representative quantum systems, showing that it also matches experimental results.
Quantum machine learning (QML) is a nascent field showing great potential in addressing complex problems. QML algorithms aim to combine the qubit’s properties, like entanglement, interference, and superposition, to perform better than any classical computation in specific tasks. This paper explores a new way of employing the Koopman operator to demonstrate its application to quantum optimization in quantum machine learning. Koopman operator-based quantum optimization has applications in various dynamical systems, where it can optimally capture the full nonlinear behavior of a system [1]. This work investigates the Koopman operator approach for quantum optimization for quantum machine learning algorithms. To do so, we capitalize on recent breakthroughs in the field, such as data-driven methods for uncovering Koopman models and the application of the Koopman operator to investigate quantum machine learning optimal strategies. Specifically, we provide a strategy-driven approach with the full pipeline for encoding classical dynamical systems into quantum-accessible representation, thus employing beneficial properties of Koopman and transfer operators. This strategy enables us to represent the state space of classical systems in terms suitable for quantum computation. Here, we introduce a framework for using the Koopman operator to represent complex, high-dimensional classical dynamics on a quantum computer to enhance classical machine learning algorithms by transferring their capabilities onto quantum systems. Additionally, we provide a programmatical framework for using the Koopman operator in quantum machine learning frameworks on the IBM Qiskit framework.
In recent years, the application of machine learning to physics has been actively explored. In this paper, we study a method for estimating the ground-state energy of quantum Hamiltonians by applying data-driven Koopman analysis within the framework of variational wave functions. Koopman theory is a framework for analyzing the nonlinear dynamics of vectors, in which the dynamics are linearized by lifting the vectors to functions defined over the original vector space. We focus on the fact that the imaginary-time Schrödinger equation, when restricted to a variational wave function, is described by a nonlinear time evolution of the variational parameter vector. We collect sample points of this nonlinear dynamics at parameter configurations where the discrepancy between the true imaginary-time dynamics and the dynamics on the variational manifold is small, and perform data-driven continuous Koopman analysis. Within our formulation, the ground-state energy is reduced to the leading eigenvalue of a differential operator known as the Koopman generator. As a concrete example, we generate samples for the four-site transverse-field Ising model and estimate the ground-state energy using extended dynamic mode decomposition (EDMD). Furthermore, as an extension of this framework, we formulate the method for the case where the variational wave function is given by a uniform matrix product state on an infinite chain. By employing computational techniques developed within the framework of the time-dependent variational principle, all the quantities required for our analysis, including error estimation, can be computed efficiently in such systems. Since our approach provides predictions for the ground-state energy even when the true ground state lies outside the variational manifold, it is expected to complement conventional variational methods.
Implementing fast and high-fidelity quantum operations using open-loop quantum optimal control relies on having an accurate model of the quantum dynamics. Any deviations between this model and the complete dynamics of the device, such as the presence of spurious modes or pulse distortions, can degrade the performance of optimal controls in practice. Here, we propose an experimentally simple approach to realize optimal quantum controls tailored to the device parameters and environment while specifically characterizing this quantum system. Concretely, we use physics-inspired machine learning to infer an accurate model of the dynamics from experimentally available data and then optimize our experimental controls on this trained model. We show the power and feasibility of this approach by optimizing arbitrary single-qubit operations on a superconducting transmon qubit, using detailed numerical simulations. We demonstrate that this framework produces an accurate description of the device dynamics under arbitrary controls, together with the precise pulses achieving arbitrary single-qubit gates with a high fidelity of about 99.99%.
The quantum harmonic oscillator is one of the most fundamental objects in physics. We consider the case where it is extended to an arbitrary number modes and includes all possible terms that are bilinear in the annihilation and creation operators, and we assume that we also have an arbitrary time-dependent drive term that is linear in those operators. Such a Hamiltonian is very general, covering a broad range of systems, including quantum optics, superconducting circuit QED, quantum error correcting codes, Bose-Einstein condensates, atomic wave packet transport beyond the adiabatic limit, and many others. We examine this situation from the point of view of quantum control, making use of optimal control theory to determine what can be accomplished, both when the controls are arbitrary and when they must minimize some cost function. In particular we develop a class of analytical pulses. We then apply our theory to a number of specific topical physical systems to illustrate its use and provide explicit control functions, including the case of the continuously driven conditional displacement gate.
Dynamic mode decomposition (DMD) is a data-driven method for the estimation, prediction, and control of complex dynamical systems, which has gained much attention in the fields of nonlinear dynamics and fluid mechanics. A DMD method for quantum spin systems, described by a linear dynamical system of a finite-dimensional set of observables, has been proposed recently. In this study, we propose two DMD methods applicable to infinite-dimensional open quantum systems, which use time-series data obtained by quantum state tomography. First, we propose a kernel DMD method for a data-driven spectral analysis of the Liouville superoperator. Second, we propose a method for the parameter estimation of the Liouville superoperator, which incorporates prior knowledge of the model structure into DMD. The proposed methods can accurately reconstruct the system dynamics and show that DMD frameworks can be applicable to infinite-dimensional open quantum systems.
Abstract Quantum optimal control (QOC) relies on accurately modeling system dynamics and is often challenged by unknown or inaccessible interactions in real systems. Taking an unknown collective spin system as an example, this work introduces a machine-learning-based, data-driven scheme to overcome the challenges encountered, with a trained neural network (NN) assuming the role of a surrogate model that captures the system’s dynamics and subsequently enables QOC to be performed on the NN instead of on the real system. The trained NN surrogate proves effective for practical QOC tasks and is further demonstrated to be adaptable to different experimental conditions, remaining robust across varying system sizes and pulse durations.