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:2202.09980 · IEEE Journal of Selected Topics in Signal Processing

Quantum-Inspired Multi-Parameter Adaptive Bayesian Estimation for Sensing and Imaging

Kwan K. Lee, Christos N. Gagatsos, Saikat Guha, Amit Ashok

Published Oct 14, 2022• 34 citations
arXivHTMLPDF

Abstract

It is well known in Bayesian estimation theory that the conditional estimator attains the minimum mean squared error (MMSE) for estimating a scalar parameter of interest. In quantum, e.g., optical and atomic, imaging and sensing tasks the user has access to the quantum state that encodes the parameter. The choice of a measurement operator, i.e. a positive-operator valued measure (POVM), leads to a measurement outcome on which the aforesaid classical MMSE estimator is employed. Personick found the optimum POVM that attains the MMSE over all possible physically allowable measurements and the resulting MMSE (Personick, 1997). This result from 1971 is less-widely known than the quantum Fisher information (QFI), which lower bounds the variance of an unbiased estimator over all measurements without considering any prior probability. For multi-parameter estimation, in quantum Fisher estimation theory the inverse of the QFI matrix provides an operator lower bound on the covariance of an unbiased estimator, and this bound is understood in the positive semidefinite sense. However, there has been little work on quantifying the quantum limits and measurement designs, for multi-parameter quantum estimation in aBayesiansetting. In this work, we build upon Personick's result to construct a Bayesian adaptive (greedy) measurement scheme for multi-parameter estimation. We illustrate our proposed measurement scheme with the application of localizing a cluster of point emitters in a highly sub-Rayleigh angular field-of-view, an important problem in fluorescence microscopy and astronomy. Our algorithm translates to a multi-spatial-mode transformation prior to a photon-detection array, with electro-optic feedback to adapt the mode sorter. We show that this receiver performs superior to quantum-noise-limited focal-plane direct imaging.

DOI
10.1109/jstsp.2022.3214774

Related papers

Bayesian quantum phase estimation with fixed photon states

Boyu Zhou, Saikat Guha, Christos N. Gagatsos

We consider a two-mode bosonic state with fixed photon number $$n \in \mathbb {N}$$ , whose upper and lower modes pick up a phase $$\phi $$ and $$-\phi $$ , respectively. We compute the optimal Fock coefficients of the input state, such that the mean square error (MSE) for estimating $$\phi $$ is minimized, while the minimum MSE is always attainable by a measurement. Our setting is Bayesian, i.e., we consider $$\phi $$ to be a random variable that follows a prior probability distribution function (PDF). Initially, we consider the flat prior PDF and we discuss the well-known fact that the MSE is not an informative tool for estimating a phase when the variance of the prior PDF is large. Therefore, we move on to study truncated versions of the flat prior in both single-shot and adaptive approaches. For our adaptive technique, we consider $$n=1$$ and truncated prior PDFs. Each subsequent step utilizes as prior PDF the posterior probability of the previous step, and at the same time we update the optimal state and optimal measurement.

arXiv:2308.01293•Quantum Information Processing•Oct 28, 2024•5 citations

Bayesian approach to simultaneous quantum multiparameter estimation with finite data

Chun Kit Dennis Law, József Zsolt Bernád

Multiparameter estimation remains a fundamental challenge in quantum estimation theory because the optimal measurements associated with different parameters are generally incompatible. In this work, we develop a Bayesian framework for the simultaneous estimation of multiple parameters. Our approach builds on Personick's method, in which the estimation problem is reduced to a set of Lyapunov equations whose solutions define optimal observables for the individual parameters. Since these observables generally do not commute, we construct a convex combination of the solutions, parameterized by a set of variational parameters. The spectral decomposition of the resulting operator defines a parametrized projection-valued measure. The resulting projective measurement determines the likelihood function, from which the minimum mean-square error estimators and the corresponding Bayesian mean-square errors are obtained, following standard Bayesian procedures, as functions of the variational parameters. To determine their optimal values, we formulate a minimax optimization problem based on the normalized Bayesian mean-square errors, introducing a min-max normalization procedure that enables a meaningful comparison of estimation errors associated with different parameters. This optimization yields a variationally optimized projective measurement for simultaneous Bayesian estimation. Two qubit examples, involving phase estimation and the estimation of parameters defining convex combinations of unitary operations, demonstrate the construction of optimized projective measurements and the corresponding minimum mean-square error estimators within the proposed framework for finite data sets.

quant-pharXiv:2608.10230•Aug 12, 2026•2 citations

Quantum limits in microscopy and spectroscopy

Evangelia Bisketzi

In recent years quantum metrology and quantum sensing have enabled the advancement of quantum technologies and research in fundamental physics. Through statistical analysis, in the context of parameter estimation, the amount of information about a parameter of interest encoded in a quantum state can be quantified. In this thesis we focus on the application of quantum estimation theory to imaging and spectroscopy with quantum light. We study the fundamental bound of the mean square error for an unbiased estimator in terms of the quantum Fisher information (QFI) for two problems, one in microscopy and one in spectroscopy. Firstly, we study the problem of localising multiple point sources below the diffraction limit. We show that localisation microscopy of multiple weak, incoherent point sources with possibly different intensities in one spatial dimension is equivalent to estimating the amplitudes of a classical mixture of coherent states of a simple harmonic oscillator. We obtain the QFI matrix elements analytically. In the regime of arbitrarily small separations we find it to be no more than rank two – implying that no more than two independent parameters can be estimated irrespective of the number of point sources. We use the eigenvalues of the classical and quantum Fisher information matrices to compare the performance of spatial-mode demultiplexing and direct imaging in localisation microscopy with respect to the quantum limits. Secondly, we study the estimation of the electric dipole moment (EDM) of a two-level atom through its interaction with quantum pulses of light in free space. We derive analytical expressions for the states of one-photon wavepackets and entangled photon pairs (EPP). We numerically calculate the QFI for different states of light, including coherent and squeezed pulses, and compare their performance in estimating the EDM. We find that the one-photon wavepacket and the EPP have similar performances, while the performance of the entangled photon pair is not improved by increasing the entanglement of the EPP. Our results indicate that using Fock states to estimate the EDM of a two-level atom is preferable to using entangled light.

•Sep 16, 2026•0 citations

Bayesian minimum mean square error for transmissivity sensing

Boyu Zhou, Boulat A. Bash, Saikat Guha +1 more

We address the problem of estimating the transmissivity of the pure-loss single-mode bosonic channel from the Bayesian point of view, i.e., when a prior probability distribution function (PDF) on the transmissivity is available. We compute the quantum limit of the Bayesian minimum mean square error of estimating transmissivity. Specifically, we consider two prior PDFs: the two-point distribution (relevant for reading from an optical drive) and the beta distribution (relevant for imaging a reflective pixelated scene). If the probe's mean photon number is an integer, for the two-point PDF we prove that the optimal probe is a Fock state and the optimal measurement is photon counting, while for the beta PDF our numerical results provide evidence on the optimality of the Fock state probe and photon counting. When the probe's mean photon number is any (nonnegative) real number, we conjecture the form of the optimal probe state and we study the performance of photon counting, which is a suboptimal yet practical measurement. Our methods can be applied for any prior PDF. We discuss how different precision metrics and priors on the parameter influence the quantum-optimal probe and measurement. To our knowledge, these results on transmissivity estimation with quantum-limited precision represent the first Bayesian analysis of the problem.

arXiv:2304.05539•Physical Review Research•Oct 11, 2023•5 citations

Quantum-limited Optical Super-resolution Imaging of Multiple Point-sources

Kwan Kit Lee, Christos N. Gagatsos, Saikat Guha +1 more

We exploit the quantum parameter estimation approach proposed by Personick [1] to develop a Bayesian adaptive sequential measurement scheme to estimate the number, brightness and location of point sources in a constellation. We demonstrate that the proposed measurement scheme outperforms the traditional diffraction-limited imaging in the deep sub-Rayleigh regime.

Imaging and Applied Optics•Jan 1, 2022•0 citations

Sufficient support size of measurements for quantum estimation

Koichi Yamagata

In quantum estimation for a $d$-parameter family of density operators on a finite-dimensional Hilbert space $\mathcal{H}$, an estimator is specified by a pair $\left(M,\hatθ\right)$, where $M$ is a POVM with a finite outcome set $Ω$ and $\hatθ:Ω\to\mathbb{R}^{d}$ is a classical estimator map. Since the number of outcomes $\left|Ω\right|$ is a priori unbounded, the space of admissible POVMs is vast, which makes the search for optimal estimators difficult. In this paper, for the minimization of the weighted trace of the mean squared error among locally unbiased estimators, we prove that it suffices to consider POVMs with at most $\left(\dim\mathcal{H}\right)^{2}+d(d+1)/2-1$ outcomes, and that the optimization can be restricted to rank-one measurements. For Bayesian estimation with a general loss function, we show that rank-one POVMs with at most $(\dim\mathcal{H})^{2}$ outcomes are sufficient for the infimum value of the Bayes risk. Furthermore, when the model admits a real sufficient subalgebra, we show that the $\left(\dim\mathcal{H}\right)^{2}$ term in the above support-size bounds can be reduced in both the locally unbiased and Bayesian settings. These bounds substantially reduce the search space for optimal measurements and justify restricting numerical optimization to rank-one POVMs with finitely many outcomes.

quant-pharXiv:2604.21323•Journal of Mathematical Physics•Apr 24, 2026•2 citations
It is well known in Bayesian estimation theory that the conditional estimator attains the minimum mean squared error (MMSE) for estimating a scalar parameter of interest. In quantum, e.g., optical and atomic, imaging and sensing tasks the user has access to the quantum state that encodes the parameter. The choice of a measurement operator, i.e. a positive-operator valued measure (POVM), leads to a measurement outcome on which the aforesaid classical MMSE estimator is employed. Personick found the optimum POVM that attains the MMSE over all possible physically allowable measurements and the resulting MMSE (Personick, 1997). This result from 1971 is less-widely known than the quantum Fisher information (QFI), which lower bounds the variance of an unbiased estimator over all measurements without considering any prior probability. For multi-parameter estimation, in quantum Fisher estimation theory the inverse of the QFI matrix provides an operator lower bound on the covariance of an unbiased estimator, and this bound is understood in the positive semidefinite sense. However, there has been little work on quantifying the quantum limits and measurement designs, for multi-parameter quantum estimation in aBayesiansetting. In this work, we build upon Personick's result to construct a Bayesian adaptive (greedy) measurement scheme for multi-parameter estimation. We illustrate our proposed measurement scheme with the application of localizing a cluster of point emitters in a highly sub-Rayleigh angular field-of-view, an important problem in fluorescence microscopy and astronomy. Our algorithm translates to a multi-spatial-mode transformation prior to a photon-detection array, with electro-optic feedback to adapt the mode sorter. We show that this receiver performs superior to quantum-noise-limited focal-plane direct imaging.
We consider a two-mode bosonic state with fixed photon number n∈Nn \in \mathbb {N}n∈N , whose upper and lower modes pick up a phase ϕ\phi ϕ and −ϕ-\phi −ϕ , respectively. We compute the optimal Fock coefficients of the input state, such that the mean square error (MSE) for estimating ϕ\phi ϕ is minimized, while the minimum MSE is always attainable by a measurement. Our setting is Bayesian, i.e., we consider ϕ\phi ϕ to be a random variable that follows a prior probability distribution function (PDF). Initially, we consider the flat prior PDF and we discuss the well-known fact that the MSE is not an informative tool for estimating a phase when the variance of the prior PDF is large. Therefore, we move on to study truncated versions of the flat prior in both single-shot and adaptive approaches. For our adaptive technique, we consider n=1n=1n=1 and truncated prior PDFs. Each subsequent step utilizes as prior PDF the posterior probability of the previous step, and at the same time we update the optimal state and optimal measurement.
Multiparameter estimation remains a fundamental challenge in quantum estimation theory because the optimal measurements associated with different parameters are generally incompatible. In this work, we develop a Bayesian framework for the simultaneous estimation of multiple parameters. Our approach builds on Personick's method, in which the estimation problem is reduced to a set of Lyapunov equations whose solutions define optimal observables for the individual parameters. Since these observables generally do not commute, we construct a convex combination of the solutions, parameterized by a set of variational parameters. The spectral decomposition of the resulting operator defines a parametrized projection-valued measure. The resulting projective measurement determines the likelihood function, from which the minimum mean-square error estimators and the corresponding Bayesian mean-square errors are obtained, following standard Bayesian procedures, as functions of the variational parameters. To determine their optimal values, we formulate a minimax optimization problem based on the normalized Bayesian mean-square errors, introducing a min-max normalization procedure that enables a meaningful comparison of estimation errors associated with different parameters. This optimization yields a variationally optimized projective measurement for simultaneous Bayesian estimation. Two qubit examples, involving phase estimation and the estimation of parameters defining convex combinations of unitary operations, demonstrate the construction of optimized projective measurements and the corresponding minimum mean-square error estimators within the proposed framework for finite data sets.
In recent years quantum metrology and quantum sensing have enabled the advancement of quantum technologies and research in fundamental physics. Through statistical analysis, in the context of parameter estimation, the amount of information about a parameter of interest encoded in a quantum state can be quantified. In this thesis we focus on the application of quantum estimation theory to imaging and spectroscopy with quantum light. We study the fundamental bound of the mean square error for an unbiased estimator in terms of the quantum Fisher information (QFI) for two problems, one in microscopy and one in spectroscopy. Firstly, we study the problem of localising multiple point sources below the diffraction limit. We show that localisation microscopy of multiple weak, incoherent point sources with possibly different intensities in one spatial dimension is equivalent to estimating the amplitudes of a classical mixture of coherent states of a simple harmonic oscillator. We obtain the QFI matrix elements analytically. In the regime of arbitrarily small separations we find it to be no more than rank two – implying that no more than two independent parameters can be estimated irrespective of the number of point sources. We use the eigenvalues of the classical and quantum Fisher information matrices to compare the performance of spatial-mode demultiplexing and direct imaging in localisation microscopy with respect to the quantum limits. Secondly, we study the estimation of the electric dipole moment (EDM) of a two-level atom through its interaction with quantum pulses of light in free space. We derive analytical expressions for the states of one-photon wavepackets and entangled photon pairs (EPP). We numerically calculate the QFI for different states of light, including coherent and squeezed pulses, and compare their performance in estimating the EDM. We find that the one-photon wavepacket and the EPP have similar performances, while the performance of the entangled photon pair is not improved by increasing the entanglement of the EPP. Our results indicate that using Fock states to estimate the EDM of a two-level atom is preferable to using entangled light.
We address the problem of estimating the transmissivity of the pure-loss single-mode bosonic channel from the Bayesian point of view, i.e., when a prior probability distribution function (PDF) on the transmissivity is available. We compute the quantum limit of the Bayesian minimum mean square error of estimating transmissivity. Specifically, we consider two prior PDFs: the two-point distribution (relevant for reading from an optical drive) and the beta distribution (relevant for imaging a reflective pixelated scene). If the probe's mean photon number is an integer, for the two-point PDF we prove that the optimal probe is a Fock state and the optimal measurement is photon counting, while for the beta PDF our numerical results provide evidence on the optimality of the Fock state probe and photon counting. When the probe's mean photon number is any (nonnegative) real number, we conjecture the form of the optimal probe state and we study the performance of photon counting, which is a suboptimal yet practical measurement. Our methods can be applied for any prior PDF. We discuss how different precision metrics and priors on the parameter influence the quantum-optimal probe and measurement. To our knowledge, these results on transmissivity estimation with quantum-limited precision represent the first Bayesian analysis of the problem.
In quantum estimation for a ddd-parameter family of density operators on a finite-dimensional Hilbert space H\mathcal{H}H, an estimator is specified by a pair (M,θ^)\left(M,\hatθ\right)(M,θ^), where MMM is a POVM with a finite outcome set ΩΩΩ and θ^:Ω→Rd\hatθ:Ω\to\mathbb{R}^{d}θ^:Ω→Rd is a classical estimator map. Since the number of outcomes ∣Ω∣\left|Ω\right|∣Ω∣ is a priori unbounded, the space of admissible POVMs is vast, which makes the search for optimal estimators difficult. In this paper, for the minimization of the weighted trace of the mean squared error among locally unbiased estimators, we prove that it suffices to consider POVMs with at most (dim⁡H)2+d(d+1)/2−1\left(\dim\mathcal{H}\right)^{2}+d(d+1)/2-1(dimH)2+d(d+1)/2−1 outcomes, and that the optimization can be restricted to rank-one measurements. For Bayesian estimation with a general loss function, we show that rank-one POVMs with at most (dim⁡H)2(\dim\mathcal{H})^{2}(dimH)2 outcomes are sufficient for the infimum value of the Bayes risk. Furthermore, when the model admits a real sufficient subalgebra, we show that the (dim⁡H)2\left(\dim\mathcal{H}\right)^{2}(dimH)2 term in the above support-size bounds can be reduced in both the locally unbiased and Bayesian settings. These bounds substantially reduce the search space for optimal measurements and justify restricting numerical optimization to rank-one POVMs with finitely many outcomes.