Trainee Project
Title:
Model-free control of superconducting qubit dynamics for quantum state preparation
Dates:
2025/03/01 - 2025/05/01
Supervisor(s):
Other supervisor(s):
Jean-Philippe CONDOMINES (jean-philippe.condomines@enac.fr)
Description:
Subject: Quantum state preparation is a fundamental process in quantum computing and quantum physics. It aims to manipulate a quantum system into
a specific desired state. This process consists of controlling a quantum system, such as a qubit, so that it evolves from an initial state to a target state,
while satisfying requirements for high fidelity and robustness to perturbations.
In the case of superconducting qubits, achieving these goals is often compromised by complex and poorly characterized dynamics, affected by
phenomena such as noise, unwanted couplings, and unmodeled quantum effects. Although the literature proposes several control methods for these
systems, model predictive control (MPC) is generally suitable when the system characteristics are well known, enabling optimal control (see article:
https://arxiv.org/abs/2201.05266). However, for superconducting qubits, accurate modeling of the dynamics often proves difficult due to unaccounted
physics, instrumental noise and unwanted couplings. These limitations reduce the effectiveness of model-based approaches. Model-free control (MFC)
presents itself as a promising alternative and has been widely explored in the literature by researchers such as Cédric Join from CRAN and Michel
Fliess from X (see article: https://arxiv.org/abs/1305.7085).
The aim of this internship will be to develop a model-free control (MFC) method for quantum state preparation in superconducting qubit dynamics,
using an existing model predictive control (MPC) code as a starting point. Initially, the trainee will use this code to master current control techniques
and explore the specifics of quantum state preparation. Then, he will adapt the code to implement model-free control, enabling control pulses to be
adjusted in real time, reactively to state feedback, without the need for advanced optimization.
a specific desired state. This process consists of controlling a quantum system, such as a qubit, so that it evolves from an initial state to a target state,
while satisfying requirements for high fidelity and robustness to perturbations.
In the case of superconducting qubits, achieving these goals is often compromised by complex and poorly characterized dynamics, affected by
phenomena such as noise, unwanted couplings, and unmodeled quantum effects. Although the literature proposes several control methods for these
systems, model predictive control (MPC) is generally suitable when the system characteristics are well known, enabling optimal control (see article:
https://arxiv.org/abs/2201.05266). However, for superconducting qubits, accurate modeling of the dynamics often proves difficult due to unaccounted
physics, instrumental noise and unwanted couplings. These limitations reduce the effectiveness of model-based approaches. Model-free control (MFC)
presents itself as a promising alternative and has been widely explored in the literature by researchers such as Cédric Join from CRAN and Michel
Fliess from X (see article: https://arxiv.org/abs/1305.7085).
The aim of this internship will be to develop a model-free control (MFC) method for quantum state preparation in superconducting qubit dynamics,
using an existing model predictive control (MPC) code as a starting point. Initially, the trainee will use this code to master current control techniques
and explore the specifics of quantum state preparation. Then, he will adapt the code to implement model-free control, enabling control pulses to be
adjusted in real time, reactively to state feedback, without the need for advanced optimization.
Keywords:
Quantum mechanics, MFC, MPC
Conditions:
Minimum 2-month and maximum 6-month internship
Department(s):
| Control Identification Diagnosis |
Funds:
ENAC Toulouse