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

European Signal Processing Conference (EUSIPCO)

Fast disentanglement-based blind quantum source separation and process tomography using a feedforward quantum-classical adapting structure

Yannick Deville, Alain Deville

Published Aug 1, 2016• 3 citations
Publisher

Abstract

Our recent investigations of blind quantum source separation and process tomography methods for Heisenberg-coupled quantum bits (qubits) were focused on introducing a new separation principle, based on output disentanglement. We here extend them by proposing a more advanced implementation of their cost function and optimization algorithm. This leads us to move from a feedback to a feedforward adapting block, which avoids potential issues related to feedback in quantum circuits. The number of quantum source state preparations required to blindly adapt the separating system is thus strongly decreased (roughly from 107to 104), yielding much faster adaptation.

DOI
10.1109/eusipco.2016.7760299

Related papers

Blind quantum source separation and process tomography systems for time-varying coupling

Yannick Deville, Alain Deville

Classical, i.e. non-quantum, blind source separation (BSS) methods estimate unknown source signals by using only mixed signals obtained by transferring these source signals through a mixing transform, which typically has unknown parameter values. We developed quantum versions of BSS (called BQSS) and we extended them to the blind estimation of the unknown mixing/coupling parameters, thus achieving blind quantum process tomography (BQPT). In these investigations, the mixing parameter values remained fixed over time. We here analyze configurations where they vary. We show that the operation of our previous system may then be extended to handle BQPT and simple BQSS configurations. BQSS with rapidly evolving mixtures is more complex, because the fundamental quantum no-cloning theorem entails that each quantum state available at the output of our separating system cannot be copied so as to be used as the inputs of both (i) the subsequent quantum system which exploits it and (ii) the block of our separating system which continuously uses these states to update the separating system parameter values. We avoid this issue by introducing a quite different BQSS system, which is based on a master-slave structure and performs parallel adaptation and source state restoration. The proposed approach is thus based on spin electronics for implementing the considered quantum states, and on an original control loop in the proposed master-slave structure, which requires a specific measurement procedure, due to the quantum nature of the signals to be processed. This paper therefore bridges the gap between advanced information processing functions (especially self-learning algorithms) and the physical devices, defined at the atomic level, required for implementing such computations.

IEEE International Workshop of Electronics, Control, Measurement, Signals and their Application to Mechatronics (ECMSM)•May 1, 2017•2 citations

Blind qubit state disentanglement with Quantum processing: Principle, criterion and algorithm using measurements along two directions

Yannick Deville, Alain Deville

In the framework of Blind Quantum Source Separation, we investigate the adaptation of a separating system which receives coupled quantum bit (qubit) states and processes them with quantum means in its feedforward path, to uncouple them. We propose the first separation principle which ensures that the output qubit states of this system are disentangled and that they restore the non-entangled source qubit states up to limited indeterminacies. This separation principle exploits measurements of output spin components along two directions and has some links with the non-quantum Independent Component Analysis (ICA) principle. It opens the way to various practical separation criteria and algorithms, some of which are described here.

IEEE International Conference on Acoustics Speech and Signal Processing•May 1, 2014•14 citations

The Blind Version of Quantum Process Tomography: Operating with Unknown Input Values

Yannick Deville, Alain Deville

System identification and inversion are two closely related problems, which lead to various configurations. In the quantum framework, system identification was only studied in the non-blind (or supervised) mode, i.e. with known input states (in addition to known output states). It is then called quantum process tomography. We here develop its blind (or unsupervised) version, operating with unknown (unentangled) input states (which are here not controlled by an automated loop). Our approach takes advantage of the blind quantum source separation (i.e. blind multi-qubit system inversion) methods based on disentanglement that we recently introduced. It is detailed for Heisenberg spin coupling and could be extended to other classes of processes. It estimates the parameter values of the Heisenberg Hamiltonian.

IFAC-PapersOnLine•Jul 1, 2017•9 citations

Fast Disentanglement-Based Blind Quantum Source Separation and Process Tomography: A Closed-Form Solution Using a Feedback Classical Adapting Structure

Yannick Deville, Alain Deville

Lecture notes in computer science•Jan 1, 2017•3 citations

A Quantum-feedforward and classical-feedback separating structure adapted with monodirectional measurements; Blind qubit uncoupling capability and links with ICA

Yannick Deville, Alain Deville

This paper deals with Blind Quantum Source Separation (BQSS). The proposed separating system only receives observed quantum states, each obtained after two quantum bits (qubits) were first independently initialized with values that this system does not know and then coupled according to a quantum operator involving parameter values that the separating system does not know a priori. This system aims at restoring the information contained in the initial qubit states. We previously tackled this problem by first converting the observed quantum state into data which have a classical (i.e. non-quantum) form and then only processing the latter data with classical means. We here propose a new approach, based on a separating system which uses quantum processing means (consisting of a specific unitary operator) in its feedforward path. Classical processing is only used in the feedback path of this system, to adapt its transfer function. The proposed adaptation method is based on binary-valued monodirectional output measurements and related to quantum disentanglement. We also analyze its relationships with classical Independent Component Analysis. We prove that this BQSS method partly restores the initial two-qubit states: it determines their components up to some phase and permutation indeterminacies.

Machine learning for signal processing ...•Sep 1, 2013•13 citations

STOCHASTIC QUANTUM INFORMATION PROCESSING, WITH APPLICATIONS TO BLIND QUANTUM SYSTEM IDENTIFICATION AND SOURCE SEPARATION

Yannick Deville, Alain Deville

Many methods for processing classical or quantum data are based on estimating some parameters, e.g. those of adaptive filters, artificial neural networks (including deep learning approaches), blind source separation systems or quantum gates. For classical signals, these methods include stochastic algorithms, such as stochastic gradient descent. We here first introduce a partly related, general, type of stochastic algorithms for quantum data and we prove the asymptotic efficiency of the proposed estimator. We then show the attractiveness of this stochastic approach for the quantum version of blind multiple-input multiple-output system identification and blind source separation: the resulting model estimation methods can operate with a single copy of each considered quantum state, whereas the previous methods require many copies of the same (unknown) states to be available and are thus "less blind". Numerical tests show that good performance is obtained with 104such quantum states.

Machine learning for signal processing ...•Sep 1, 2018•4 citations
Our recent investigations of blind quantum source separation and process tomography methods for Heisenberg-coupled quantum bits (qubits) were focused on introducing a new separation principle, based on output disentanglement. We here extend them by proposing a more advanced implementation of their cost function and optimization algorithm. This leads us to move from a feedback to a feedforward adapting block, which avoids potential issues related to feedback in quantum circuits. The number of quantum source state preparations required to blindly adapt the separating system is thus strongly decreased (roughly from 107to 104), yielding much faster adaptation.
Classical, i.e. non-quantum, blind source separation (BSS) methods estimate unknown source signals by using only mixed signals obtained by transferring these source signals through a mixing transform, which typically has unknown parameter values. We developed quantum versions of BSS (called BQSS) and we extended them to the blind estimation of the unknown mixing/coupling parameters, thus achieving blind quantum process tomography (BQPT). In these investigations, the mixing parameter values remained fixed over time. We here analyze configurations where they vary. We show that the operation of our previous system may then be extended to handle BQPT and simple BQSS configurations. BQSS with rapidly evolving mixtures is more complex, because the fundamental quantum no-cloning theorem entails that each quantum state available at the output of our separating system cannot be copied so as to be used as the inputs of both (i) the subsequent quantum system which exploits it and (ii) the block of our separating system which continuously uses these states to update the separating system parameter values. We avoid this issue by introducing a quite different BQSS system, which is based on a master-slave structure and performs parallel adaptation and source state restoration. The proposed approach is thus based on spin electronics for implementing the considered quantum states, and on an original control loop in the proposed master-slave structure, which requires a specific measurement procedure, due to the quantum nature of the signals to be processed. This paper therefore bridges the gap between advanced information processing functions (especially self-learning algorithms) and the physical devices, defined at the atomic level, required for implementing such computations.
In the framework of Blind Quantum Source Separation, we investigate the adaptation of a separating system which receives coupled quantum bit (qubit) states and processes them with quantum means in its feedforward path, to uncouple them. We propose the first separation principle which ensures that the output qubit states of this system are disentangled and that they restore the non-entangled source qubit states up to limited indeterminacies. This separation principle exploits measurements of output spin components along two directions and has some links with the non-quantum Independent Component Analysis (ICA) principle. It opens the way to various practical separation criteria and algorithms, some of which are described here.
System identification and inversion are two closely related problems, which lead to various configurations. In the quantum framework, system identification was only studied in the non-blind (or supervised) mode, i.e. with known input states (in addition to known output states). It is then called quantum process tomography. We here develop its blind (or unsupervised) version, operating with unknown (unentangled) input states (which are here not controlled by an automated loop). Our approach takes advantage of the blind quantum source separation (i.e. blind multi-qubit system inversion) methods based on disentanglement that we recently introduced. It is detailed for Heisenberg spin coupling and could be extended to other classes of processes. It estimates the parameter values of the Heisenberg Hamiltonian.
This paper deals with Blind Quantum Source Separation (BQSS). The proposed separating system only receives observed quantum states, each obtained after two quantum bits (qubits) were first independently initialized with values that this system does not know and then coupled according to a quantum operator involving parameter values that the separating system does not know a priori. This system aims at restoring the information contained in the initial qubit states. We previously tackled this problem by first converting the observed quantum state into data which have a classical (i.e. non-quantum) form and then only processing the latter data with classical means. We here propose a new approach, based on a separating system which uses quantum processing means (consisting of a specific unitary operator) in its feedforward path. Classical processing is only used in the feedback path of this system, to adapt its transfer function. The proposed adaptation method is based on binary-valued monodirectional output measurements and related to quantum disentanglement. We also analyze its relationships with classical Independent Component Analysis. We prove that this BQSS method partly restores the initial two-qubit states: it determines their components up to some phase and permutation indeterminacies.
Many methods for processing classical or quantum data are based on estimating some parameters, e.g. those of adaptive filters, artificial neural networks (including deep learning approaches), blind source separation systems or quantum gates. For classical signals, these methods include stochastic algorithms, such as stochastic gradient descent. We here first introduce a partly related, general, type of stochastic algorithms for quantum data and we prove the asymptotic efficiency of the proposed estimator. We then show the attractiveness of this stochastic approach for the quantum version of blind multiple-input multiple-output system identification and blind source separation: the resulting model estimation methods can operate with a single copy of each considered quantum state, whereas the previous methods require many copies of the same (unknown) states to be available and are thus "less blind". Numerical tests show that good performance is obtained with 104such quantum states.