© 2026 Aqora Quantum S.A.S.

Terms Privacy Legal Notice

arXiv:1806.04503

Tutorial: Maximum likelihood estimation in the context of an optical measurement

Anthony Vella

Published Jun 8, 2018• 4 citations

Abstract

The method of maximum likelihood estimation (MLE) is a widely used statistical approach for estimating the values of one or more unknown parameters of a probabilistic model based on observed data. In this tutorial, I briefly review the mathematical foundations of MLE, then reformulate the problem for the measurement of a spatially-varying optical intensity distribution. In this context, the detection of each individual photon is treated as a random event, the outcome being the photon's location. A typical measurement consists of many detected photons, which accumulate to form a spatial intensity profile. Here, I show a straightforward derivation for the likelihood function and Fisher information matrix (FIM) associated with a measurement of multiple photons incident on a detector comprised of a discrete array of pixels. An estimate for the parameter(s) of interest may then be obtained by maximizing the likelihood function, while the FIM determines the uncertainty of the estimate. To illustrate these concepts, several simple examples are presented for the one- and two-parameter cases, revealing many interesting properties of the MLE formalism, as well as some practical considerations for optical experiments. Throughout these examples, connections are also drawn to optical applications of quantum weak measurements, including off-null ellipsometry and scatterometry.

DOI: 10.48550/arxiv.1806.04503

Related papers

Joint probability of instants of occurrence of photon events and estimation of optical parameters

Bahaa E. A. Saleh

Using a Karhunen-Loeve modal expansion of a stationary and gaussian optical field an expression is derived for the joint probability of instants of occurrence of photon events. When the observation time is much longer than the coherence time, this expression takes a simple form. In this case, and for lorentzian light, it is shown that measurement of the total number of counts is sufficient to determine the maximum likelihood estimate (MLE) of the average count rate. A nonlinear equation is found whose solution gives the MLE of the bandwidth.

Journal of Physics A Mathematical Nuclear and General•Jul 21, 1974•2 citations

Multiparameter Maximum Information States for Coherent Diffraction Measurements

Bram Verreussel, Jacob Seifert, Allard P. Mosk

In metrology, Fisher information is an important metric that quantifies the precision that can be achieved in a measurement. For optical measurements using coherent light it has been shown that Fisher information can be expressed simply using the scattering matrix of the system. Fisher information can be maximized over the input modes to achieve maximum information states, which produce optimally precise estimates for a parameter when the system is limited by photon noise. Here, we extend this approach to multiparameter estimation, in which case Fisher information takes the form of a matrix. We consider several scalar functions of the Fisher matrix to optimize the precision in multiple parameters at the same time. We also consider strategies for dealing with nuisance parameters, which can degrade the achievable precision of other parameters but are not of interest to measure. We corroborate our findings numerically using a scattering system of 2D coupled dipoles.

arXiv:2606.01445•Sep 8, 2026•0 citations

Maximum-likelihood estimation of quantum measurement

Jaromı́r Fiurášek

Maximum-likelihood estimation is applied to the determination of an unknown quantum measurement. The calibrated measuring apparatus carries out measurements on many different quantum states and the positive operator-valued measure governing the measurement statistics is then inferred from the collected data via the maximum-likelihood principle. In contrast to the procedures based on linear inversion, our approach always provides a physically sensible result. We illustrate the method on the case of the Stern-Gerlach apparatus.

arXiv:quant-ph/0101027•Physical Review A•Jul 17, 2001•175 citations

Maximum-likelihood method in quantum estimation

Paris, M G A, D'Ariano, G M, Sacchi, M F

The maximum-likelihood method for quantum estimation is reviewed and applied to the reconstruction of density matrix of spin and radiation as well as to the determination of several parameters of interest in quantum optics.

•Jan 1, 2001•16 citations

Maximum-likelihood method in quantum estimation

M. G. A. Paris

The maximum-likelihood method for quantum estimation is reviewed and applied to the reconstruction of density matrix of spin and radiation as well as to the determination of several parameters of interest in quantum optics.

arXiv:quant-ph/0101071•AIP conference proceedings•Jan 1, 2001•12 citations

Fisher information, disorder, and optical signal estimation

B. Roy Frieden

Optical signals may be regarded as probability laws (on photon events). The problem of optical signal recovery may therefore be viewed as a problem in estimating a probability law p(x). Maximum entropy has, in the past, been used for this purpose. Here, we discuss instead the use of minimum Fisher information I = ∫ d x [ p ′ ( x ) ] 2 / p ( x ) . Consider a physical system consisting of a random aggregate of particles or pho tons. Consider the experiment of measuring one particle's coordinate y, and from this, best estimating the mean coordinate θ of the ensemble, where y = θ + x. As time passes, the particles randomly become more spread out, so that the error in the estimate should increase. But the error goes inversely as I. Therefore, I should statistically decrease with time. With no constraint on the system, as t → ∞, I → 0, defining a maximally random law p(x) = constant. However, in the presence of a physical constraint, I should approach a finite value, obeying I = minimum. When the constraint is linear in the mean kinetic energy of the system, the solution p(x) obeys the stationary differential equation for the system. In this way, the Schrodinger (energy) wave equation, Helmholtz wave equation, diffusion equation, and Maxwell-Boltzmann law may be derived from one classical principle of disorder.

Annual Meeting Optical Society of America•Jan 1, 1989•0 citations

The method of maximum likelihood estimation (MLE) is a widely used statistical approach for estimating the values of one or more unknown parameters of a probabilistic model based on observed data. In this tutorial, I briefly review the mathematical foundations of MLE, then reformulate the problem for the measurement of a spatially-varying optical intensity distribution. In this context, the detection of each individual photon is treated as a random event, the outcome being the photon's location. A typical measurement consists of many detected photons, which accumulate to form a spatial intensity profile. Here, I show a straightforward derivation for the likelihood function and Fisher information matrix (FIM) associated with a measurement of multiple photons incident on a detector comprised of a discrete array of pixels. An estimate for the parameter(s) of interest may then be obtained by maximizing the likelihood function, while the FIM determines the uncertainty of the estimate. To illustrate these concepts, several simple examples are presented for the one- and two-parameter cases, revealing many interesting properties of the MLE formalism, as well as some practical considerations for optical experiments. Throughout these examples, connections are also drawn to optical applications of quantum weak measurements, including off-null ellipsometry and scatterometry.

Using a Karhunen-Loeve modal expansion of a stationary and gaussian optical field an expression is derived for the joint probability of instants of occurrence of photon events. When the observation time is much longer than the coherence time, this expression takes a simple form. In this case, and for lorentzian light, it is shown that measurement of the total number of counts is sufficient to determine the maximum likelihood estimate (MLE) of the average count rate. A nonlinear equation is found whose solution gives the MLE of the bandwidth.

In metrology, Fisher information is an important metric that quantifies the precision that can be achieved in a measurement. For optical measurements using coherent light it has been shown that Fisher information can be expressed simply using the scattering matrix of the system. Fisher information can be maximized over the input modes to achieve maximum information states, which produce optimally precise estimates for a parameter when the system is limited by photon noise. Here, we extend this approach to multiparameter estimation, in which case Fisher information takes the form of a matrix. We consider several scalar functions of the Fisher matrix to optimize the precision in multiple parameters at the same time. We also consider strategies for dealing with nuisance parameters, which can degrade the achievable precision of other parameters but are not of interest to measure. We corroborate our findings numerically using a scattering system of 2D coupled dipoles.

Maximum-likelihood estimation is applied to the determination of an unknown quantum measurement. The calibrated measuring apparatus carries out measurements on many different quantum states and the positive operator-valued measure governing the measurement statistics is then inferred from the collected data via the maximum-likelihood principle. In contrast to the procedures based on linear inversion, our approach always provides a physically sensible result. We illustrate the method on the case of the Stern-Gerlach apparatus.

Optical signals may be regarded as probability laws (on photon events). The problem of optical signal recovery may therefore be viewed as a problem in estimating a probability law p(x). Maximum entropy has, in the past, been used for this purpose. Here, we discuss instead the use of minimum Fisher information I = ∫ d x [ p ′ ( x ) ] 2 / p ( x ) . Consider a physical system consisting of a random aggregate of particles or pho tons. Consider the experiment of measuring one particle's coordinate y, and from this, best estimating the mean coordinate θ of the ensemble, where y = θ + x. As time passes, the particles randomly become more spread out, so that the error in the estimate should increase. But the error goes inversely as I. Therefore, I should statistically decrease with time. With no constraint on the system, as t → ∞, I → 0, defining a maximally random law p(x) = constant. However, in the presence of a physical constraint, I should approach a finite value, obeying I = minimum. When the constraint is linear in the mean kinetic energy of the system, the solution p(x) obeys the stationary differential equation for the system. In this way, the Schrodinger (energy) wave equation, Helmholtz wave equation, diffusion equation, and Maxwell-Boltzmann law may be derived from one classical principle of disorder.