Fisher-information training of optical sensing front ends from natural fluctuations
Abstract
Fisher-information-based optimization of programmable optical front ends in situ generally requires a response model or parameter-labeled measurements. We show that natural parameter fluctuations can instead provide the training signal. We construct proxies for the Fisher information (FI) of one parameter and the prior-weighted FI spectrum of multiple parameters from measured covariances after subtracting conditional detection noise, assuming locally affine responses. We use simulated photon counts from two incoherent point sources to train a mode sorter that approaches the quantum limit for separation and the Nagaoka--Hayashi single-copy bound for joint centroid--separation estimation. We demonstrate count-only updates that require no instrument response model and achieve performance comparable to model-based gradients. Our results provide a route to adaptive optical measurements without controlled parameter scans.
Related
- Quantum Information-Theoretic Analysis of Optical Fiber Sensing via Stimulated Brillouin Scattering
- Sensing quantum vacuum fluctuations with non-Gaussian electronic noise
- Stochastic and quantum autonomous information machine: efficiency fluctuations, precision and optimization
- Propagation of pulse fluctuations of multimode sources in optical fibers and waveguides with arbitrary chromatic dispersion
- Chern--Simons Fluctuations and Information Geometry in Discrete Electromagnetism
BibTeX
@misc{sinha2026fisherinformation,
title = {{Fisher-information training of optical sensing front ends from natural fluctuations}},
author = {Abhinav Sinha and Kai Wang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.24915},
eprint = {2609.24915},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.24915},
}
RIS
TY - UNPB
TI - Fisher-information training of optical sensing front ends from natural fluctuations
AU - Abhinav Sinha
AU - Kai Wang
PY - 2026
DA - 2026/09/21
PB - arXiv
DO - 10.48550/arxiv.2609.24915
UR - https://arxiv.org/abs/2609.24915
AB - Fisher-information-based optimization of programmable optical front ends in situ generally requires a response model or parameter-labeled measurements. We show that natural parameter fluctuations can instead provide the training signal. We construct proxies for the Fisher information (FI) of one parameter and the prior-weighted FI spectrum of multiple parameters from measured covariances after subtracting conditional detection noise, assuming locally affine responses. We use simulated photon counts from two incoherent point sources to train a mode sorter that approaches the quantum limit for separation and the Nagaoka--Hayashi single-copy bound for joint centroid--separation estimation. We demonstrate count-only updates that require no instrument response model and achieve performance comparable to model-based gradients. Our results provide a route to adaptive optical measurements without controlled parameter scans.
ER -