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:2101.01190 · Machine Learning Science and Technology

Control of stochastic quantum dynamics by differentiable programming

Frank Schäfer, Pavel Sekatski, Martin Koppenhöfer, Christoph Bruder, Michal Kloc

Published Mar 4, 2021• 20 citations
arXivHTMLPDF

Abstract

Abstract Control of the stochastic dynamics of a quantum system is indispensable in fields such as quantum information processing and metrology. However, there is no general ready-made approach to the design of efficient control strategies. Here, we propose a framework for the automated design of control schemes based on differentiable programming. We apply this approach to the state preparation and stabilization of a qubit subjected to homodyne detection. To this end, we formulate the control task as an optimization problem where the loss function quantifies the distance from the target state, and we employ neural networks (NNs) as controllers. The system’s time evolution is governed by a stochastic differential equation (SDE). To implement efficient training, we backpropagate the gradient information from the loss function through the SDE solver using adjoint sensitivity methods. As a first example, we feed the quantum state to the controller and focus on different methods of obtaining gradients. As a second example, we directly feed the homodyne detection signal to the controller. The instantaneous value of the homodyne current contains only very limited information on the actual state of the system, masked by unavoidable photon-number fluctuations. Despite the resulting poor signal-to-noise ratio, we can train our controller to prepare and stabilize the qubit to a target state with a mean fidelity of around 85%. We also compare the solutions found by the NN to a hand-crafted control strategy.

DOI
10.1088/2632-2153/abec22

Related papers

A differentiable programming method for quantum control

Frank Schäfer, Michal Kloc, Christoph Bruder +1 more

Abstract Optimal control is highly desirable in many current quantum systems, especially to realize tasks in quantum information processing. We introduce a method based on differentiable programming to leverage explicit knowledge of the differential equations governing the dynamics of the system. In particular, a control agent is represented as a neural network that maps the state of the system at a given time to a control pulse. The parameters of this agent are optimized via gradient information obtained by direct differentiation through both the neural network and the differential equation of the system. This fully differentiable reinforcement learning approach ultimately yields time-dependent control parameters optimizing a desired figure of merit. We demonstrate the method’s viability and robustness to noise in eigenstate preparation tasks for three systems: a single qubit, a chain of qubits, and a quantum parametric oscillator.

arXiv:2002.08376•Machine Learning Science and Technology•May 29, 2020•50 citations

Optimal control of stochastic Schrödinger equation via physics informed neural networks: applications in quantum optics

S Vijayaram, Pagavathigounder Balasubramaniam

In this paper, authors present a physics-informed neural networks (PINN) approach for solving the optimal control problem (OCP) of the stochastic Schrödinger equation (SSE), with an emphasis on applications in quantum optics. The SSE models quantum systems under uncertainty, a scenario often encountered in open quantum systems where environmental interactions introduce randomness into the system dynamics. These stochastic dynamics are particularly relevant in quantum optics, where precise control over quantum states is crucial for applications such as quantum computing, secure communication, and measurement protocols. Traditional numerical methods for solving such problems face limitations due to high computational costs and the complexity associated with stochastic partial differential equations. The proposed approach utilises PINN, a type of deep learning model that integrates physical laws directly into the learning process by incorporating loss functions based on governing equations, initial and boundary conditions. The fidelity-based objective function is constructed to achieve the desired quantum state so as to develop an OCP solving algorithm. The proposed work is distinguished between deterministic drift-based training and post-training stochastic validation. Extensive numerical experiments are demonstrated near-unity deterministic fidelity. Further, a systematic architecture study validated that a moderately deep network yielded optimal performance with reduced computational cost. Fidelity dissipation analysis confirmed that the method maintained accuracy even under strong damping. These findings established the PINN as an efficient and physically consistent framework for quantum optimal control in open systems.

The Philosophical Magazine A Journal of Theoretical Experimental and Applied Physics•Jun 2, 2026•0 citations

Pointwise and dynamic programming control synthesis for finite-level open quantum memory systems

Igor G. Vladimirov, Ian R. Petersen, Guodong Shi

This paper is concerned with finite-level quantum memory systems for retaining initial dynamic variables in the presence of external quantum noise. The system variables have an algebraic structure, similar to that of the Pauli matrices, and their Heisenberg picture evolution is governed by a quasilinear quantum stochastic differential equation. The latter involves a Hamiltonian whose parameters depend affinely on a classical control signal in the form of a deterministic function of time. The memory performance is quantified by a mean-square deviation of quantum system variables of interest from their initial conditions. We relate this functional to a matrix-valued state of an auxiliary classical control-affine dynamical system. This leads to a pointwise control design where the control signal minimises the time-derivative of the mean-square deviation with an additional quadratic penalty on the control. In an alternative finite-horizon setting with a terminal-integral cost functional, we apply dynamic programming and obtain a quadratically nonlinear Hamilton-Jacobi-Bellman equation, for which a solution is outlined in the form of a recursively computed asymptotic expansion.

math.OCCross-listedarXiv:2603.29225•Apr 1, 2026•0 citations

Learning quantum-state feedback control with backpropagation-free stochastic optimization

Ethan N. Evans, Ziyi Wang, Adam G. Frim +2 more

High-fidelity state preparation represents a fundamental challenge in the application of quantum technology. While the majority of optimal control approaches use feedback to improve the controller, the controller itself often does not incorporate explicit state dependence. Here, we present a general framework for training deep feedback networks for open quantum systems with continuous weak measurement that allows a variety of system and control structures that are prohibitive by many other techniques and can in effect react to unmodeled effects through nonlinear filtering. Our approach benefits from characteristics of both stochastic sampling and gradient-based optimization methods yet does not require differentiability as in backpropagation approaches. We demonstrate that this method is efficient due to inherent parallelizability, robust to open system interactions, and outperforms landmark state-dependent feedback control results in simulation.

Physical Review A•Nov 3, 2022•1 citation

Quantum-Inspired Stochastic Differential Equations for Reinforcement Learning

Jincheng Zhang

This paper introduces a novel approach to reinforcement learning (RL) utilizing stochastic differential equations (SDEs) inspired by quantum mechanics. Traditional SDEs employed in RL frequently suffer from computational complexity and reliance on approximation techniques. This work proposes a framework that directly incorporates principles of quantum measurement and superposition to generate SDEs characterized by intrinsic stochasticity. The underlying mechanism leverages the inherent uncertainty present in quantum measurements to drive policy optimization, potentially leading to more efficient and accurate learning. We formulate the SDEs using a representation that mimics quantum state evolution, introducing a novel operator that allows for controlled stochastic noise generation. The theoretical framework is presented, outlining the key equations and their implications for policy learning. The approach demonstrates the potential to bypass classical approximations inherent in existing SDE-based RL methods, offering a promising direction for enhanced performance and scalability. The core of the work lies in the reformulation of the stochastic differential equation, shifting from a purely deterministic approach to one that reflects the probabilistic nature of quantum systems. This results in an SDE that is fundamentally more amenable to efficient sampling and optimization.

•Aug 28, 2026•0 citations

A Neural Network Approach to Sampling Based Learning Control for Quantum System with Uncertainty

Jingshi Li, Song Chen Song Chen, Yanzhao Cao +1 more

Robust control design for quantum systems with uncertainty is a key task for developing practical quantum technology. In this paper, we apply neural networks to learn the control of a quantum system with uncertainty. By exploiting the auto differentiation function developed for neural network models, our method avoids the manual computation of the gradient of the cost function as required in traditional methods. We implement our method using two algorithms. One uses neural networks to learn both the states and the controls and one uses neural networks to learn only the controls but solve the states by finite difference methods. Both algorithms incorporate the sampling-based learning process into the training of the networks. The performance of the algorithms is evaluated on a practical numerical example, followed by a detailed discussion about the advantage and trade-offs between our method and the other numerical schemes.

Communications in Computational Physics•Oct 25, 2021•2 citations
Abstract Control of the stochastic dynamics of a quantum system is indispensable in fields such as quantum information processing and metrology. However, there is no general ready-made approach to the design of efficient control strategies. Here, we propose a framework for the automated design of control schemes based on differentiable programming. We apply this approach to the state preparation and stabilization of a qubit subjected to homodyne detection. To this end, we formulate the control task as an optimization problem where the loss function quantifies the distance from the target state, and we employ neural networks (NNs) as controllers. The system’s time evolution is governed by a stochastic differential equation (SDE). To implement efficient training, we backpropagate the gradient information from the loss function through the SDE solver using adjoint sensitivity methods. As a first example, we feed the quantum state to the controller and focus on different methods of obtaining gradients. As a second example, we directly feed the homodyne detection signal to the controller. The instantaneous value of the homodyne current contains only very limited information on the actual state of the system, masked by unavoidable photon-number fluctuations. Despite the resulting poor signal-to-noise ratio, we can train our controller to prepare and stabilize the qubit to a target state with a mean fidelity of around 85%. We also compare the solutions found by the NN to a hand-crafted control strategy.
Abstract Optimal control is highly desirable in many current quantum systems, especially to realize tasks in quantum information processing. We introduce a method based on differentiable programming to leverage explicit knowledge of the differential equations governing the dynamics of the system. In particular, a control agent is represented as a neural network that maps the state of the system at a given time to a control pulse. The parameters of this agent are optimized via gradient information obtained by direct differentiation through both the neural network and the differential equation of the system. This fully differentiable reinforcement learning approach ultimately yields time-dependent control parameters optimizing a desired figure of merit. We demonstrate the method’s viability and robustness to noise in eigenstate preparation tasks for three systems: a single qubit, a chain of qubits, and a quantum parametric oscillator.
In this paper, authors present a physics-informed neural networks (PINN) approach for solving the optimal control problem (OCP) of the stochastic Schrödinger equation (SSE), with an emphasis on applications in quantum optics. The SSE models quantum systems under uncertainty, a scenario often encountered in open quantum systems where environmental interactions introduce randomness into the system dynamics. These stochastic dynamics are particularly relevant in quantum optics, where precise control over quantum states is crucial for applications such as quantum computing, secure communication, and measurement protocols. Traditional numerical methods for solving such problems face limitations due to high computational costs and the complexity associated with stochastic partial differential equations. The proposed approach utilises PINN, a type of deep learning model that integrates physical laws directly into the learning process by incorporating loss functions based on governing equations, initial and boundary conditions. The fidelity-based objective function is constructed to achieve the desired quantum state so as to develop an OCP solving algorithm. The proposed work is distinguished between deterministic drift-based training and post-training stochastic validation. Extensive numerical experiments are demonstrated near-unity deterministic fidelity. Further, a systematic architecture study validated that a moderately deep network yielded optimal performance with reduced computational cost. Fidelity dissipation analysis confirmed that the method maintained accuracy even under strong damping. These findings established the PINN as an efficient and physically consistent framework for quantum optimal control in open systems.
This paper is concerned with finite-level quantum memory systems for retaining initial dynamic variables in the presence of external quantum noise. The system variables have an algebraic structure, similar to that of the Pauli matrices, and their Heisenberg picture evolution is governed by a quasilinear quantum stochastic differential equation. The latter involves a Hamiltonian whose parameters depend affinely on a classical control signal in the form of a deterministic function of time. The memory performance is quantified by a mean-square deviation of quantum system variables of interest from their initial conditions. We relate this functional to a matrix-valued state of an auxiliary classical control-affine dynamical system. This leads to a pointwise control design where the control signal minimises the time-derivative of the mean-square deviation with an additional quadratic penalty on the control. In an alternative finite-horizon setting with a terminal-integral cost functional, we apply dynamic programming and obtain a quadratically nonlinear Hamilton-Jacobi-Bellman equation, for which a solution is outlined in the form of a recursively computed asymptotic expansion.
High-fidelity state preparation represents a fundamental challenge in the application of quantum technology. While the majority of optimal control approaches use feedback to improve the controller, the controller itself often does not incorporate explicit state dependence. Here, we present a general framework for training deep feedback networks for open quantum systems with continuous weak measurement that allows a variety of system and control structures that are prohibitive by many other techniques and can in effect react to unmodeled effects through nonlinear filtering. Our approach benefits from characteristics of both stochastic sampling and gradient-based optimization methods yet does not require differentiability as in backpropagation approaches. We demonstrate that this method is efficient due to inherent parallelizability, robust to open system interactions, and outperforms landmark state-dependent feedback control results in simulation.
This paper introduces a novel approach to reinforcement learning (RL) utilizing stochastic differential equations (SDEs) inspired by quantum mechanics. Traditional SDEs employed in RL frequently suffer from computational complexity and reliance on approximation techniques. This work proposes a framework that directly incorporates principles of quantum measurement and superposition to generate SDEs characterized by intrinsic stochasticity. The underlying mechanism leverages the inherent uncertainty present in quantum measurements to drive policy optimization, potentially leading to more efficient and accurate learning. We formulate the SDEs using a representation that mimics quantum state evolution, introducing a novel operator that allows for controlled stochastic noise generation. The theoretical framework is presented, outlining the key equations and their implications for policy learning. The approach demonstrates the potential to bypass classical approximations inherent in existing SDE-based RL methods, offering a promising direction for enhanced performance and scalability. The core of the work lies in the reformulation of the stochastic differential equation, shifting from a purely deterministic approach to one that reflects the probabilistic nature of quantum systems. This results in an SDE that is fundamentally more amenable to efficient sampling and optimization.
Robust control design for quantum systems with uncertainty is a key task for developing practical quantum technology. In this paper, we apply neural networks to learn the control of a quantum system with uncertainty. By exploiting the auto differentiation function developed for neural network models, our method avoids the manual computation of the gradient of the cost function as required in traditional methods. We implement our method using two algorithms. One uses neural networks to learn both the states and the controls and one uses neural networks to learn only the controls but solve the states by finite difference methods. Both algorithms incorporate the sampling-based learning process into the training of the networks. The performance of the algorithms is evaluated on a practical numerical example, followed by a detailed discussion about the advantage and trade-offs between our method and the other numerical schemes.