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

ACS Photonics

Closed-Loop, Event-Driven Machine Learning for Autonomous Triage of Single-Photon Emitters

Tanvir M. Mahim, Md Nahin Islam, M. Mosaddequr Rahman, Abu S. M. Mohsin

Published Oct 15, 2026• 0 citations
Publisher

Abstract

Abstract Identifying which candidate sites on a solid-state sample are single-photon emitters requires a Hanbury Brown–Twiss (HBT) measurement of the second-order correlation function g(2)(τ), which is slow because photon coincidences accumulate at a rate proportional to the square of the detected count rate. We examine how much of this time can be saved by deciding while the coincidences are still accumulating and by adapting the exposure of each site. A spiking neural network reads the coincidence record in short time slices and updates, after every slice, its estimate of the probability that the site hosts a single emitter; a reinforcement-learning controller decides, site by site, whether to expose longer, certify, or reject; and a graph that encodes the electronic level structure of each emitter class conditions one estimator on the class being measured. All components are trained on simulated data from a standard three-level photon-counting model, which we validate against the exact master-equation solution before use. In simulation, the event-driven estimator reaches 90% balanced accuracy with 6 times less acquisition time than a multistart least-squares fit, typically committing to a decision 312 ms into a 1 s exposure, and adaptive exposure screens simulated 48-site fields 1.7 times faster than fixed-dwell scanning at matched precision and recall. Applied without retraining to published experimental HBT data from an InGaAs/GaAs quantum dot, a network trained only on simulated pulsed-excitation histograms estimates g(2)(0) from 30 s records with half the error of peak-area analysis of the same records. The closed-loop results are obtained in simulation only. The code, data, and trained models are openly available.

DOI
10.1021/acsphotonics.6c01984

Related papers

Physics-Based versus Data-Driven Classification of Single-Photon Quantum Emitters from Sparse Autocorrelation Data

Nhat Minh Nguyen, Md Shakhawath Hossain, Duc Anh Ngo +4 more

Identifying single photon emitters from large, inhomogeneous candidate populations is key to realizing many quantum applications. This requires measuring the emitters second order autocorrelation function, whose statistical reliability is fundamentally limited by acquisition time. Machine learning classifiers can accelerate identification from sparse data, but their performance relative to physics-based inference has not been systematically examined. Here, we introduce sequential Bayesian inference for single photon emitter classification and benchmark it against Levenberg-Marquardt fitting and a feedforward neural network. We use synthetic training and test data calibrated against real Hanbury Brown-Twiss measurements from hexagonal boron nitride emitters, enabling evaluation against an exactly known ground-truth emitter number under realistic noise and background conditions. All three approaches achieve high, near-perfect accuracy with sufficient integration time, but differ greatly in convergence rate and robustness under sparse photon statistics. The neural network is most robust at short integration times. The Bayesian classifier reaches near-perfect accuracy fastest, while retaining full physical interpretability. Levenberg-Marquardt fitting remains a valuable, fully interpretable method, achieving the highest recall despite being the slowest to converge. These results lead to several key conclusions. No single method dominates across all performance metrics. Relying on any one metric alone can give a misleading picture of classifier performance, particularly under sparse photon statistics. Physics-based and data-driven methods are complementary rather than competing approaches. Together, these findings provide practical guidance for selecting and combining classification strategies for scalable single-photon-source screening and other quantum-emitter characterization tasks.

quant-pharXiv:2608.19528•Aug 21, 2026•0 citations

Experimental certification of level dynamics in single-photon emitters

Lukáš Lachman, Ilya P. Radko, Maxime Bergamin +3 more

Emitters of single photons are essential resources for emerging quantum technologies and developed within different platforms including nonlinear optics and atomic and solid-state systems. The energy-level structures of emission processes are critical for reaching and controlling high-quality sources. The most commonly applied test uses a Hanbury Brown and Twiss (HBT) setup to determine the emitter energy-level structure based on fitting temporal correlations of photon detection events. However, only partial information about the emission process is extracted from such detection, that might be followed by an inconclusive fitting of the data. This process predetermines our limited ability to quantify and understand the dynamics in the photon emission process that are of importance for the applications in communication, sensing, and computing. In this work, we present a complete analysis based on all normalized coincidences between detection and no-detection events recorded in the same HBT setup to certify expected properties of an emitted photonic state. As a proof of concept we apply our methodology to single nitrogen-vacancy centers in diamond, in which case the certification conclusively rejects a model based on a two-level emitter that radiates a photonic state mixed with any classical noise background. Published by the American Physical Society 2024

arXiv:2407.13273•Physical Review Research•Sep 5, 2024•2 citations

Adversarial autoencoder ensemble for fast and probabilistic reconstructions of few-shot photon correlation functions for solid-state quantum emitters

Andrew H. Proppe, Kin Long Kelvin Lee, Cristian L. Cortes +11 more

Second-order photon correlation measurements [${g}^{(2)}(\ensuremath{\tau})$ functions] are widely used to classify single-photon emission purity in quantum emitters or to measure the multiexciton quantum yield of emitters that can simultaneously host multiple excitations -- such as quantum dots -- by evaluating the value of ${g}^{(2)}(\ensuremath{\tau}=0)$. Accumulating enough photons to accurately calculate this value is time consuming and could be accelerated by fitting of few-shot photon correlations. Here, we develop an uncertainty-aware, deep adversarial autoencoder ensemble (AAE) that reconstructs noise-free ${g}^{(2)}(\ensuremath{\tau})$ functions from noise-dominated, few-shot inputs. The model is trained with simulated ${g}^{(2)}(\ensuremath{\tau})$ functions that are facilely generated by Poisson sampling time bins. The AAE reconstructions are performed orders-of-magnitude faster, with reconstruction errors and estimates of ${g}^{(2)}(\ensuremath{\tau}=0)$ that are lower in variance and similar in accuracy compared to Maximum likelihood estimation and Levenberg-Marquardt least-squares fitting approaches, for simulated and experimentally measured few-shot ${g}^{(2)}(\ensuremath{\tau})$ functions (\ensuremath{\sim}100 two-photon events) of InP/ZnS/ZnSe and CdS/CdSe/CdS quantum dots. The deep-ensemble model comprises eight individual autoencoders, allowing for probabilistic reconstructions of noise-free ${g}^{(2)}(\ensuremath{\tau})$ functions, and we show that the predicted variance scales inversely with number of shots, with comparable uncertainties to computationally intensive Markov chain Monte Carlo sampling. This work demonstrates the advantage of machine learning models to perform uncertainty-aware, fast, and accurate reconstructions of simple Poisson-distributed photon correlation functions, allowing for on-the-fly reconstructions and accelerated materials characterization of solid-state quantum emitters.

Physical review. B./Physical review. B•Jul 29, 2022•14 citations

Observation of photon antibunching with only one standard single-photon detector

Shaojie Liu, Xing Lin, Feng Liu +3 more

The second-order photon correlation function g2(τ) is of great importance in quantum optics. g2(τ) is typically measured with the Hanbury Brown and Twiss (HBT) interferometer, which employs a pair of single-photon detectors and a dual-channel time acquisition module. Here, we demonstrate a new method to measure and extract g2(τ) with a standard single-photon avalanche photodiode (dead-time = 22 ns) and a single-channel time acquisition module. This is realized by shifting the coincidence counts of interest to a time window not affected by the dead-time and after-pulse of the detection system using a fiber-based delay line. The new scheme is verified by measuring g2(τ) from a single colloidal nanocrystal. Photon antibunching is unambiguously observed and agrees well with the result measured using the standard HBT setup. Our scheme simplifies the higher-order correlation technique and might be favored in cost-sensitive circumstances.

arXiv:2008.12669•Review of Scientific Instruments•Jan 1, 2021•3 citations

Transfer learning and the early estimation of single-photon source quality using machine learning methods

David Jacob Kedziora, Anna Musiał, Wojciech Rudno-Rudziński +1 more

Abstract The use of single-photon sources (SPSs) is central to numerous systems and devices proposed amidst a modern surge in quantum technology. However, manufacturing schemes remain imperfect, and single-photon emission purity must often be experimentally verified via interferometry. Such a process is typically slow and costly, which has motivated growing research into whether SPS quality can be more rapidly inferred from incomplete emission statistics. Hence, this study is a sequel to previous work that demonstrated significant uncertainty in the standard method of quality estimation, i.e. the least-squares fitting of a physically motivated function, and asks: can machine learning (ML) do better? The study leverages eight datasets obtained from measurements involving an exemplary quantum emitter, i.e. a single InGaAs/GaAs epitaxial quantum dot; these eight contexts predominantly vary in the intensity of the exciting laser. Specifically, via a form of ‘transfer learning’, five ML models, three linear and two ensemble-based, are trained on data from seven of the contexts and tested on the eighth. Validation metrics quickly reveal that even a linear regressor can outperform standard fitting when it is tested on the same contexts it was trained on, but the success of transfer learning is less assured, even though statistical analysis, made possible by data augmentation, suggests its superiority as an early estimator. Accordingly, the study concludes by discussing future strategies for grappling with the problem of SPS context dissimilarity, e.g. feature engineering and model adaptation.

arXiv:2408.11322•Machine Learning Science and Technology•Apr 2, 2025•2 citations

Qualification and Simulation of a Single Photon Source

Máté Galambos, Petra Seres, Eszter Udvary

Single-photon sources are key components of a quantum communication network; however, present-day realizations of such photon sources do not yet show an ideal behavior. A quantitative way to qualify them is by calculating the second order correlation function, as done in the Hanbury Brown-Twiss (HBT) experiment [1]. In this study a Bell-source that produces an entangled Bell sate was used as a single-photon source. When one branch is used as heralding for the other the source produces sub-Poisson photon statistics. This is not true for the non-heralded case. While conducting the HBT experiment in the non-heralded case we came across a phenomenon of double coincidence peaks. Through our study we found that this could be a result of two very distinct phenomenon, one of which is due to the non-ideal behavior of a single-photon source, the other due to secondary photons generated by single-photon avalanche diodes (SPADs). Both an ideal sub-Poisson statistic and the double coincidence peaks result in g(2)(0) = 0 by the HBT experiment, but in the first case the photon source is indeed an actual single-photon source, while in the second case this could also happen with a non-ideal single-photon source. For that reason, sometimes it is not sufficient to rely only on the HBT experiment to qualify a single-photon source, but also conduct some deeper experiments.

International Conference on Transparent Optical Networks•Jul 6, 2025•0 citations
Abstract Identifying which candidate sites on a solid-state sample are single-photon emitters requires a Hanbury Brown–Twiss (HBT) measurement of the second-order correlation function g(2)(τ), which is slow because photon coincidences accumulate at a rate proportional to the square of the detected count rate. We examine how much of this time can be saved by deciding while the coincidences are still accumulating and by adapting the exposure of each site. A spiking neural network reads the coincidence record in short time slices and updates, after every slice, its estimate of the probability that the site hosts a single emitter; a reinforcement-learning controller decides, site by site, whether to expose longer, certify, or reject; and a graph that encodes the electronic level structure of each emitter class conditions one estimator on the class being measured. All components are trained on simulated data from a standard three-level photon-counting model, which we validate against the exact master-equation solution before use. In simulation, the event-driven estimator reaches 90% balanced accuracy with 6 times less acquisition time than a multistart least-squares fit, typically committing to a decision 312 ms into a 1 s exposure, and adaptive exposure screens simulated 48-site fields 1.7 times faster than fixed-dwell scanning at matched precision and recall. Applied without retraining to published experimental HBT data from an InGaAs/GaAs quantum dot, a network trained only on simulated pulsed-excitation histograms estimates g(2)(0) from 30 s records with half the error of peak-area analysis of the same records. The closed-loop results are obtained in simulation only. The code, data, and trained models are openly available.
Identifying single photon emitters from large, inhomogeneous candidate populations is key to realizing many quantum applications. This requires measuring the emitters second order autocorrelation function, whose statistical reliability is fundamentally limited by acquisition time. Machine learning classifiers can accelerate identification from sparse data, but their performance relative to physics-based inference has not been systematically examined. Here, we introduce sequential Bayesian inference for single photon emitter classification and benchmark it against Levenberg-Marquardt fitting and a feedforward neural network. We use synthetic training and test data calibrated against real Hanbury Brown-Twiss measurements from hexagonal boron nitride emitters, enabling evaluation against an exactly known ground-truth emitter number under realistic noise and background conditions. All three approaches achieve high, near-perfect accuracy with sufficient integration time, but differ greatly in convergence rate and robustness under sparse photon statistics. The neural network is most robust at short integration times. The Bayesian classifier reaches near-perfect accuracy fastest, while retaining full physical interpretability. Levenberg-Marquardt fitting remains a valuable, fully interpretable method, achieving the highest recall despite being the slowest to converge. These results lead to several key conclusions. No single method dominates across all performance metrics. Relying on any one metric alone can give a misleading picture of classifier performance, particularly under sparse photon statistics. Physics-based and data-driven methods are complementary rather than competing approaches. Together, these findings provide practical guidance for selecting and combining classification strategies for scalable single-photon-source screening and other quantum-emitter characterization tasks.
Emitters of single photons are essential resources for emerging quantum technologies and developed within different platforms including nonlinear optics and atomic and solid-state systems. The energy-level structures of emission processes are critical for reaching and controlling high-quality sources. The most commonly applied test uses a Hanbury Brown and Twiss (HBT) setup to determine the emitter energy-level structure based on fitting temporal correlations of photon detection events. However, only partial information about the emission process is extracted from such detection, that might be followed by an inconclusive fitting of the data. This process predetermines our limited ability to quantify and understand the dynamics in the photon emission process that are of importance for the applications in communication, sensing, and computing. In this work, we present a complete analysis based on all normalized coincidences between detection and no-detection events recorded in the same HBT setup to certify expected properties of an emitted photonic state. As a proof of concept we apply our methodology to single nitrogen-vacancy centers in diamond, in which case the certification conclusively rejects a model based on a two-level emitter that radiates a photonic state mixed with any classical noise background. Published by the American Physical Society 2024
Second-order photon correlation measurements [${g}^{(2)}(\ensuremath{\tau})$ functions] are widely used to classify single-photon emission purity in quantum emitters or to measure the multiexciton quantum yield of emitters that can simultaneously host multiple excitations -- such as quantum dots -- by evaluating the value of ${g}^{(2)}(\ensuremath{\tau}=0)$. Accumulating enough photons to accurately calculate this value is time consuming and could be accelerated by fitting of few-shot photon correlations. Here, we develop an uncertainty-aware, deep adversarial autoencoder ensemble (AAE) that reconstructs noise-free ${g}^{(2)}(\ensuremath{\tau})$ functions from noise-dominated, few-shot inputs. The model is trained with simulated ${g}^{(2)}(\ensuremath{\tau})$ functions that are facilely generated by Poisson sampling time bins. The AAE reconstructions are performed orders-of-magnitude faster, with reconstruction errors and estimates of ${g}^{(2)}(\ensuremath{\tau}=0)$ that are lower in variance and similar in accuracy compared to Maximum likelihood estimation and Levenberg-Marquardt least-squares fitting approaches, for simulated and experimentally measured few-shot ${g}^{(2)}(\ensuremath{\tau})$ functions (\ensuremath{\sim}100 two-photon events) of InP/ZnS/ZnSe and CdS/CdSe/CdS quantum dots. The deep-ensemble model comprises eight individual autoencoders, allowing for probabilistic reconstructions of noise-free ${g}^{(2)}(\ensuremath{\tau})$ functions, and we show that the predicted variance scales inversely with number of shots, with comparable uncertainties to computationally intensive Markov chain Monte Carlo sampling. This work demonstrates the advantage of machine learning models to perform uncertainty-aware, fast, and accurate reconstructions of simple Poisson-distributed photon correlation functions, allowing for on-the-fly reconstructions and accelerated materials characterization of solid-state quantum emitters.
The second-order photon correlation function g2(τ) is of great importance in quantum optics. g2(τ) is typically measured with the Hanbury Brown and Twiss (HBT) interferometer, which employs a pair of single-photon detectors and a dual-channel time acquisition module. Here, we demonstrate a new method to measure and extract g2(τ) with a standard single-photon avalanche photodiode (dead-time = 22 ns) and a single-channel time acquisition module. This is realized by shifting the coincidence counts of interest to a time window not affected by the dead-time and after-pulse of the detection system using a fiber-based delay line. The new scheme is verified by measuring g2(τ) from a single colloidal nanocrystal. Photon antibunching is unambiguously observed and agrees well with the result measured using the standard HBT setup. Our scheme simplifies the higher-order correlation technique and might be favored in cost-sensitive circumstances.
Abstract The use of single-photon sources (SPSs) is central to numerous systems and devices proposed amidst a modern surge in quantum technology. However, manufacturing schemes remain imperfect, and single-photon emission purity must often be experimentally verified via interferometry. Such a process is typically slow and costly, which has motivated growing research into whether SPS quality can be more rapidly inferred from incomplete emission statistics. Hence, this study is a sequel to previous work that demonstrated significant uncertainty in the standard method of quality estimation, i.e. the least-squares fitting of a physically motivated function, and asks: can machine learning (ML) do better? The study leverages eight datasets obtained from measurements involving an exemplary quantum emitter, i.e. a single InGaAs/GaAs epitaxial quantum dot; these eight contexts predominantly vary in the intensity of the exciting laser. Specifically, via a form of ‘transfer learning’, five ML models, three linear and two ensemble-based, are trained on data from seven of the contexts and tested on the eighth. Validation metrics quickly reveal that even a linear regressor can outperform standard fitting when it is tested on the same contexts it was trained on, but the success of transfer learning is less assured, even though statistical analysis, made possible by data augmentation, suggests its superiority as an early estimator. Accordingly, the study concludes by discussing future strategies for grappling with the problem of SPS context dissimilarity, e.g. feature engineering and model adaptation.
Single-photon sources are key components of a quantum communication network; however, present-day realizations of such photon sources do not yet show an ideal behavior. A quantitative way to qualify them is by calculating the second order correlation function, as done in the Hanbury Brown-Twiss (HBT) experiment [1]. In this study a Bell-source that produces an entangled Bell sate was used as a single-photon source. When one branch is used as heralding for the other the source produces sub-Poisson photon statistics. This is not true for the non-heralded case. While conducting the HBT experiment in the non-heralded case we came across a phenomenon of double coincidence peaks. Through our study we found that this could be a result of two very distinct phenomenon, one of which is due to the non-ideal behavior of a single-photon source, the other due to secondary photons generated by single-photon avalanche diodes (SPADs). Both an ideal sub-Poisson statistic and the double coincidence peaks result in g(2)(0) = 0 by the HBT experiment, but in the first case the photon source is indeed an actual single-photon source, while in the second case this could also happen with a non-ideal single-photon source. For that reason, sometimes it is not sufficient to rely only on the HBT experiment to qualify a single-photon source, but also conduct some deeper experiments.