Controlling the Flow of Information in Optical Metrology
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: We're looking at the paper: "Controlling the Flow of Information in Optical Metrology".
Mira: Here's the gist of the paper. Optical metrology has progressed beyond the Abbe–Rayleigh limit by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations Here, we show that Fisher information can be viewed as a physical entity that propagates through space, and we derive a wave equation for sensitivity fields describing its flow This perspective reframes metrology as a discipline in which resolution can be engineered by tailoring information sources and flow for applications in atomic-scale diagnostics and beyond The Conceptual Framework of Information Flow Fisher information, which defines the fundamental precision limit, is conceptualized as a physical entity that propagates through space This interpretation suggests that Fisher information is conserved in dissipation-free regions, much like energy The flow of Fisher information inherits the full wave-like character of light itself, allowing it to resonate, diffract, and interfere This phenomenon leads to a new paradigm for optical design where information is treated as a wave-like entity to be actively guided and controlled through interference and choice of materials Fisher Information and Scattering The creation of Fisher information in a scattering process crucially depends on the resonant nature of the scattering target The maximum amount of Fisher information one can extract through an optimal measurement procedure is equal to the quantum Fisher information and is given by J = 4(dθcsaat)/ħω Resonances originating from Mie or plasmonic scattering emerge as pronounced peaks in this expression The proportionality factor in J proportional to kPscat for a single multipolar resonance depends strongly on the probe beam's polarization with respect to the parameter of interest Anisotropic Information Transport The ratio of Fisher information on displacements along x:y:z is found to be 1:2:7 when the incident plane wave is polarized along x This strong dependence suggests that simply choosing the right incoming polarization can provide a remarkable advantage in a measurement process For higher order resonances, the electric and magnetic quadrupoles give the same Fisher information regardless of polarization Engineering Information Control The wave-like nature of Fisher information is demonstrated by its propagation governed by a wave equation for the sensitivity fields This allows for the manipulation of information flow through structures like gratings and resonators The Information Talbot effect shows that grating structures provide a simple means of controlling not only physical field patterns but also sensitivity fields and thus the flow of Fisher information Reciprocity and Optimization A reciprocity relation is established for sensitivity fields, which provides a symmetry principle for information exchange in scattering systems This framework enables computationally efficient inverse design of structures that tailor and optimise Fisher information flow, such as directional radiating antennas The adjoint variable method extends naturally to objectives involving sensitivity fields, relying on the symmetry of the underlying differential operator The work demonstrates that Fisher information is a physically propagating, wave-like quantity governed by electromagnetic principles By tailoring the target geometry and its surrounding environment, researchers can enhance, suppress, or redirect this information flow with high precision This framework allows optical metrology to be reimagined as an exercise in “information field engineering” The findings open avenues for the development of “information photonics” where the physical structure of information flow is a design target The metrology of the future can become both more precise and more adaptive, with applications in nanotechnology, biomedical imaging, and quantum sensing.
Kai: Pruning redundant reasoning tokens The paper is 2508.13640 The single most informative sentence about the paper's subject and finding is Optical metrology has progressed beyond the Abbe–Rayleigh limit by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations.
Mira: How the training signal is built The authors show how Fisher information and the scattered power behave for different wavelengths using the equation ∂theta2(csaat†csaat) = 2(dθcsaat†dθcsaat) + 2Re((dθ2csaat†csaat)) This leads to the expression J = 2(ħω−1)(dθ/dPscat) − 4(ħω−1)Re((dθ2csaat†csaat)) Where the gains come from The authors identify the effective sources of Fisher information as arising from spatial regions where optical fields are sensitive to variations in the measured parameter These give rise to sensitivity fields that can interfere, and under certain conditions, destructively suppress each other, leading to cancellations in the information flow Tailoring the target geometry and its surrounding environment allows for the enhancement, suppression, or redirection of this information flow with high precision.
Kai: The authors also demonstrate how reciprocity manifests for sensitivity fields, providing a symmetry principle for information exchange in scattering systems This framework enables computationally efficient inverse design of structures that tailor and optimise Fisher information flow.
Title and authors: Kai: Before we get into what they actually found, I just want to talk about the title and who wrote this paper, "Controlling the Flow of Information in Optical Metrology".
Mira: It’s interesting because it immediately suggests an active control mechanism rather than just a passive measurement technique.
Lev: I mean, when you read that title, you expect some sort of engineering application where you can manipulate the input to get a better output.
Kai: Right. The authors are Weimar, Zhou, Neubacher, Grant, Hüpfl, MacDonald and Rotter. They’re all from institutes in Vienna and Southampton.
Mira: It's a good mix of theoretical physics grounding and experimental work in optics and sensing technology across different universities.
Lev: That’s what I like to see—a team that bridges the gap between the pure theory of how information flows and the actual hardware you have to build for it.
The paper's summary: Kai: Okay, so this paper explains that Fisher information defines our fundamental precision limit, but instead of just calculating it, they model it as a physical entity that propagates through space.
Mira: They show how this information is conserved in areas where there’s no energy loss, like in dissipation-free regions. This means the flow of Fisher information has the same wave-like character as light itself.
Lev: So they're saying this information can resonate, diffract, and interfere with other fields, which opens up new ways to design optical systems beyond simple beam steering.
Kai: They then focus on how material composition and the geometry of the target dictate where this information is generated and how it travels toward our detector.
Mira: That’s the key part: they reveal that you can steer this flow analogously to how you steer a light beam, using specific designs to enhance or suppress it.
Lev: It sounds like they’re moving from just measuring a static point to designing the entire environment around the measurement to shape the result.
The paper's improvements: Kai: The paper points out that Fisher information can be resonantly enhanced in structured environments, which is a big deal for improving resolution.
Mira: And they show how you can steer this flow using structures like gratings and resonators, showing an Information Talbot effect where those structures control both the physical field pattern and the sensitivity fields.
Lev: That ties directly into what we need for hardware; if we can design a structure that steers the information flow, it becomes a functional part of the sensing element itself.
Kai: They also established a reciprocity relation for these sensitivity fields, which acts like a symmetry principle for how information exchanges happen in scattering systems.
Mira: That symmetry allows for computationally efficient inverse design of structures, like directional radiating antennas, by using something called the adjoint variable method.
Lev: So it means instead of simulating every possible shape to find the best one, you can use this symmetry to optimize the structure much faster for a specific information flow direction.
Conclusion: Kai: To wrap up, the main point of "Controlling the Flow of Information in Optical Metrology" is that Fisher information is a physical wave quantity governed by electromagnetic principles.
Mira: By tailoring what happens around the target, we can enhance or suppress this information flow with high precision, which redefines metrology as "information field engineering."
Lev: From my side, what I see is that the reciprocity framework they developed offers a practical way to make inverse design computationally feasible for directional sensing hardware.
Kai: So this work opens up avenues for designing instruments where the physical structure of information flow is a direct design target, moving toward information photonics.
Mira: It suggests a future where metrology is not just about getting the best measurement, but actively shaping the information landscape itself for applications like nanotechnology and quantum sensing.
Lev: It sets a path for designing directional antennas and other components that sculpt electromagnetic energy precisely based on these flow dynamics.
Maximilian Weimar, Huanli Zhou, Luca Neubacher, Thomas A. Grant, Jakob Hüpfl, Kevin F. MacDonald, Stefan Rotter, Nikolay I. Zheludev
Institute for Theoretical Physics, Vienna University of Technology (TU Wien) · Optoelectronics Research Centre, University of Southampton · Hagler Institute for Advanced Study, Texas A&M University
physics.optics
Submitted: 2025-08-19
Updated: 2026-02-10
Comments: 28 pages, 7 figures
Journal ref: Nature Materials (2026)
DOI: 10.1038/s41563-026-02789-y
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 85/100
The gist: Optical metrology has progressed beyond the Abbe–Rayleigh limit by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations Here…
Key concepts
- Fisher Information as a Physical Entity
- Fisher information, which sets the fundamental limit on measurement precision, is conceptualized as a physical entity that travels through space. It behaves like energy in dissipation-free regions. This wave-like nature allows it to resonate and interfere with other fields, providing a new way to design optical systems by treating information as something that can be actively guided and controlled.
- Wave Equation for Sensitivity Fields
- The sensitivity fields describe how Fisher information flows through space, governed by a wave equation. This framework allows researchers to manipulate this flow using structures like gratings and resonators. By controlling the physical structure, one can directly control the flow of information itself.
- Anisotropic Information Transport
- The way Fisher information is transported depends heavily on polarization and direction. For example, when polarized along x, the ratio of information on x:y:z displacements is 1:2:7. This demonstrates that choosing the correct incoming polarization can give a significant advantage in measurement precision.
- Information Field Engineering
- This is the new paradigm where optical metrology is viewed as an exercise in engineering the flow of information. By tailoring target geometries and their surroundings, researchers can enhance, suppress, or redirect this information with high precision. The goal is to sculpt the spatial and spectral distribution of information.
Terminology
Summary
Optical metrology has progressed beyond the Abbe–Rayleigh limit by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations Here, we show that Fisher information can be viewed as a physical entity that propagates through space, and we derive a wave equation for sensitivity fields describing its flow This perspective reframes metrology as a discipline in which resolution can be engineered by tailoring information sources and flow for applications in atomic-scale diagnostics and beyond
The Conceptual Framework of Information Flow
Fisher information, which defines the fundamental precision limit, is conceptualized as a physical entity that propagates through space This interpretation suggests that Fisher information is conserved in dissipation-free regions, much like energy The flow of Fisher information inherits the full wave-like character of light itself, allowing it to resonate, diffract, and interfere This phenomenon leads to a new paradigm for optical design where information is treated as a wave-like entity to be actively guided and controlled through interference and choice of materials
Fisher Information and Scattering
The creation of Fisher information in a scattering process crucially depends on the resonant nature of the scattering target The maximum amount of Fisher information one can extract through an optimal measurement procedure is equal to the quantum Fisher information and is given by J = 4(dθcsaat)/ħω Resonances originating from Mie or plasmonic scattering emerge as pronounced peaks in this expression The proportionality factor in J proportional to kPscat for a single multipolar resonance depends strongly on the probe beam's polarization with respect to the parameter of interest
Anisotropic Information Transport
The ratio of Fisher information on displacements along x:y:z is found to be 1:2:7 when the incident plane wave is polarized along x This strong dependence suggests that simply choosing the right incoming polarization can provide a remarkable advantage in a measurement process For higher order resonances, the electric and magnetic quadrupoles give the same Fisher information regardless of polarization
Engineering Information Control
The wave-like nature of Fisher information is demonstrated by its propagation governed by a wave equation for the sensitivity fields This allows for the manipulation of information flow through structures like gratings and resonators The Information Talbot effect shows that grating structures provide a simple means of controlling not only physical field patterns but also sensitivity fields and thus the flow of Fisher information
Reciprocity and Optimization
A reciprocity relation is established for sensitivity fields, which provides a symmetry principle for information exchange in scattering systems This framework enables computationally efficient inverse design of structures that tailor and optimise Fisher information flow, such as directional radiating antennas The adjoint variable method extends naturally to objectives involving sensitivity fields, relying on the symmetry of the underlying differential operator
Conclusion
The work demonstrates that Fisher information is a physically propagating, wave-like quantity governed by electromagnetic principles By tailoring the target geometry and its surrounding environment, researchers can enhance, suppress, or redirect this information flow with high precision This framework allows optical metrology to be reimagined as an exercise in “information field engineering” The findings open avenues for the development of “information photonics” where the physical structure of information flow is a design target The metrology of the future can become both more precise and more adaptive, with applications in nanotechnology, biomedical imaging, and quantum sensing.
Pruning redundant reasoning tokens
The paper is 2508.13640 The single most informative sentence about the paper's subject and finding is Optical metrology has progressed beyond the Abbe–Rayleigh limit by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations.
How the training signal is built
The authors show how Fisher information and the scattered power behave for different wavelengths using the equation ∂theta2(csaat†csaat) = 2(dθcsaat†dθcsaat) + 2Re((dθ2csaat†csaat)) This leads to the expression J = 2(ħω−1)(dθ/dPscat) − 4(ħω−1)Re((dθ2csaat†csaat))
Where the gains come from
The authors identify the effective sources of Fisher information as arising from spatial regions where optical fields are sensitive to variations in the measured parameter These give rise to sensitivity fields that can interfere, and under certain conditions, destructively suppress each other, leading to cancellations in the information flow Tailoring the target geometry and its surrounding environment allows for the enhancement, suppression, or redirection of this information flow with high precision. The authors also demonstrate how reciprocity manifests for sensitivity fields, providing a symmetry principle for information exchange in scattering systems This framework enables computationally efficient inverse design of structures that tailor and optimise Fisher information flow. p. 8 This extension relies on the symmetry of the underlying
differential operator, which governs the propagation of both the physical fields and the associated
sensitivity fields. As a result, this framework shall enable computationally efficient inverse design
of structures that tailor and optimise Fisher information flow, such as directional radiating
antennas. In the same way as the reciprocity of
many scattering systems is routinely used to reduce the complexity of physical measurements and
numerical simulations, similar considerations can be applied to Fisher information. For instance,
inverse design processes with Fisher information-based objectives, such as maximising or
controlling information flow along certain spatial directions, will benefit from reciprocity: In the
Methods section, we show how the widely used adjoint-variable method32,33 extends naturally to
objectives involving sensitivity fields. In the same way as the reciprocity of
many scattering systems is routinely used to reduce the complexity of physical measurements and
numerical simulations, similar considerations can be applied to Fisher information. For instance,
inverse design processes with Fisher information-based objectives, such as maximising or
controlling information flow along certain spatial directions, will benefit from reciprocity: In the
Methods section, we show how the widely used adjoint-variable method32,33 extends naturally to
objectives involving sensitivity fields.Where the paper says it
Where the paper says it
Where the paper says it
The authors identify the effective sources of Fisher information as arising from spatial regions where optical fields are sensitive to variations in the measured parameter These give rise to sensitivity fields that can interfere, and under certain conditions, destructively suppress each other, leading to cancellations in the information flow Tailoring the target geometry and its surrounding environment allows for the enhancement, suppression, or redirection of this information flow with high precision The authors also demonstrate how reciprocity manifests for sensitivity fields, providing a symmetry principle for information exchange in scattering systems This framework enables computationally efficient inverse design of structures that tailor and optimise Fisher information flow. The final goal is to construct instruments to shape the spatial and spectral distribution of information, similar to how antennas are used to sculpt electromagnetic energy.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Continuous flow of Fisher information modeling allows AI systems to be designed as
information field engineers
capable of actively guiding and controlling resolution by tailoring information sources and flow through interference and choice of materials. This enables instruments to be constructed to shape the spatial and spectral distribution of information, similar to how antennas sculpt electromagnetic energy. -
Real-time tracking capabilities are enhanced by using neural networks as measurement estimators trained on variations, achieving sub-atomic length scales beyond 100 picometers with up to a million measurements per second for thermal motion tracking. This allows AI systems to perform ultra-precise real-time localization of nanoscale objects in dynamic environments.
-
AI systems can be optimized for directional sensing by exploiting polarization dependence, such as using incident polarization to maximize Fisher information on specific displacement parameters (e.g., 1:2:7 ratio for x:y:z displacements). This allows the AI to select optimal measurement setups based on expected environmental conditions to maximize precision.
-
Inverse design processes for (nano)photonic structures can be performed more efficiently by using adjoint variable methods extended to sensitivity fields, which enables the optimization of structures that tailor and optimize Fisher information flow along certain spatial directions. This facilitates computationally efficient inverse design of directional radiating antennas.
-
Metrology instruments can be designed to utilize superoscillatory fields to achieve sub-diffractive localization scales, potentially surpassing the conventional Abbe diffraction limit for resolution in imaging applications. This capability allows AI systems to surpass traditional imaging limitations by leveraging near-field enhancement and control of information flow via superoscillatory lenses.
Abstract
Optical metrology has progressed beyond the Abbe-Rayleigh limit, unlocking (sub)atomic precision by leveraging nonlinear phenomena, statistical accumulation, and AI estimators trained on measurand variations. Here, we show that Fisher information, which defines the fundamental precision limit, can be viewed as a physical entity that propagates through space, and we derive a wave equation for sensitivity fields describing its flow, which can resonate, diffract, and interfere. We reveal how material composition, geometry, and environmental design dictate where information is generated and how it travels, analogous to antennas and metasurfaces sculpting electromagnetic energy. Plasmonic and dielectric resonances enhance information flow, while gratings and near-field structures reshape radiation patterns. This perspective reframes metrology as a discipline in which resolution can be engineered by tailoring information sources and flow for applications in atomic-scale diagnostics and beyond, including optimisation of Light Detection and Ranging (LiDAR), remote sensing, and radar technologies.
Related papers
- Raman optical activity and helicity-dependent phonon splitting in chiral CoSi
- Stability and optical quality of "windmill"-formed 8CB liquid crystal films for replenishable plasma mirrors
- Experimental Scaling of Diffraction Efficiency in Laser-Induced Plasma Gratings
- Photonic Exceptional Topological Bound States in the Continuum
- SHAPE: a sparse Hamiltonian approach for prescribing eigenstates in metamaterials
- All-photonic millimeter-wave and terahertz signal core at room and cryogenic temperatures