Abstract Single quantum light emitters are valuable resources for engineered quantum systems. They can function as robust single-photon generators, allow optical control of single spins, provide readout capabilities for atomic-scale sensors, and provide interfaces between stationary and flying qubits. Environmental factors can lead to single emitters exhibiting ‘blinking’, whereby the fluorescence level switches between on and off states. Detailed characterisation of this blinking behaviour including determining the switching rates is often a powerful way to gain understanding about the underlying physical mechanisms. While simple thresholds can be used to identify the on and off intervals and thus extract the rates from the time-series of counts for bright emitters with low background noise, such approaches become difficult for emitters fluorescing at low levels, high noise, or switching at fast rates. We develop a Bayesian approach capable of inferring switching rates directly from the time-series. This is able to deal with high levels of noise and fast switching in fluorescence traces. Moreover, the Bayesian inference also yields a robust picture of the parameter uncertainties, providing a benefit also for bright emitters in low-noise settings. The technique can be adapted to identify the underlying states as well as extracting the rates of switching. Finally, our method is applicable to a broad range of systems that show behaviour analogous to a blinking emitter.
Abstract Single quantum light emitters are valuable resources for engineered quantum systems. They can function as robust single-photon generators, allow optical control of single spins, provide readout capabilities for atomic-scale sensors, and provide interfaces between stationary and flying qubits. Environmental factors can lead to single emitters exhibiting ‘blinking’, whereby the fluorescence level switches between on and off states. Detailed characterisation of this blinking behaviour including determining the switching rates is often a powerful way to gain understanding about the underlying physical mechanisms. While simple thresholds can be used to identify the on and off intervals and thus extract the rates from the time-series of counts for bright emitters with low background noise, such approaches become difficult for emitters fluorescing at low levels, high noise, or switching at fast rates. We develop a Bayesian approach capable of inferring switching rates directly from the time-series. This is able to deal with high levels of noise and fast switching in fluorescence traces. Moreover, the Bayesian inference also yields a robust picture of the parameter uncertainties, providing a benefit also for bright emitters in low-noise settings. The technique can be adapted to identify the underlying states as well as extracting the rates of switching. Finally, our method is applicable to a broad range of systems that show behaviour analogous to a blinking emitter.
ABSTRACT To unfold their ultimate performance, a large number of photonics‐based quantum technologies require reliable emitters of single photons. Emission intensity fluctuations in these single‐photon emitters occur over a broad range of timescales from nanoseconds to more than several milliseconds, depending on the emitter, posing significant challenges for their future applicability. This paper focuses on quantum emitters with intensity fluctuations that occur at a rate comparable to the detection rate of the emitted light. In this regime, there is approximately one photon per bright period. Our method enables the determination of the bright‐state intensity, which is otherwise unobservable through a direct time‐domain measurement. This method is validated through simulations in various scenarios, and an analytical model is derived. We apply the method to measurements of both quantum emitters in hexagonal boron nitride and of centers in diamond. The gain in information about the states contributing to the emission dynamics enhances the understanding of bunching phenomena and, in particular, clarifies the influence of potential background light. This is crucial for identifying, designing, and fabricating bright single‐photon emitters for applications in quantum technologies.
Quantum state readout with minimal resources is crucial for scalable quantum information processing. As a leading platform, neutral atom arrays rely on atomic fluorescence imaging for qubit readout, requiring short exposure, low photon count schemes to mitigate heating and atom loss while enabling mid-circuit feedback. However, a fundamental challenge arises in the single-photon regime where severe overlap in state distributions causes conventional threshold discrimination to fail. Here, we report an AI-accelerated Bayesian inference method for fluorescence readout in neutral atom arrays. Our approach leverages Bayesian inference to achieve reliable state detection at the single-photon level under short exposure. Specifically, we introduce a weakly anchored Bayesian scheme that requires calibration of only one state, addressing asymmetric calibration challenges common across quantum platforms. Furthermore, acceleration is achieved via a permutation-invariant neural network, which yields a 100-fold speedup by compressing iterative inference into a single forward pass. The approach achieves relative readout fidelity above 99% and 98% for histogram overlaps of 61% and 72%, respectively, enabling reliable extraction of Rabi oscillations and Ramsey interference results unattainable with conventional threshold based methods. This framework supports scalable, real-time readout of large atom arrays and paves the way toward AI-enhanced quantum technology in computation and sensing.
The blinking statistics of quantum emitters and their corresponding Markov models play an important role in high-resolution microscopy of biological samples as well as in nano-optoelectronics and many other fields of science and engineering. Current methods for analyzing the blinking statistics like the full counting statistics and the Viterbi algorithm break down for low photon rates. We present an evaluation scheme that eliminates the need for both a minimum photon flux and the usual binning of photon events which limits the measurement bandwidth. Our approach is based on higher-order spectra of the measurement record which we model within the recently introduced method of quantum polyspectra from the theory of continuous quantum measurements. By virtue of this approach we can determine on- and off-switching rates of a semiconductor quantum dot at light levels 1000 times lower than in a standard experiment and 20 times lower those than achieved with a scheme from full counting statistics. Thus, a very powerful high-bandwidth approach to the parameter-learning task of single-photon hidden Markov models is established with applications in many fields of science.
Abstract Realisation of experiments even on small and medium-scale quantum computers requires an optimisation of several parameters to achieve high-fidelity operations. As the size of the quantum register increases, the characterisation of quantum states becomes more difficult since the number of parameters to be measured grows as well and finding efficient observables in order to estimate the parameters of the model becomes a crucial task. Here we propose a method relying on application of Bayesian inference that can be used to determine systematic, unknown phase shifts of multi-qubit states. This method offers important advantages as compared to Ramsey-type protocols. First, application of Bayesian inference allows the selection of an adaptive basis for the measurements which yields the optimal amount of information about the phase shifts of the state. Secondly, this method can process the outcomes of different observables at the same time. This leads to a substantial decrease in the resources needed for the estimation of phases, speeding up the state characterisation and optimisation in experimental implementations. The proposed Bayesian inference method can be applied in various physical platforms that are currently used as quantum processors.
This thesis intends to familiarize the reader with the concepts of photon statistics and correlations in quantum optics. Developing light sources that emit quantum states is central for the realization of quantum technologies. One important step in characterizing these sources is the measurement of field fluctuations and correlations, by coincidence measurements. The expectation value of a coincidence measurement, a simultaneous measurement of two intensities (or, more general, four fields), is represented by the fourth-order correlation function. The value of the correlation function, at zero delay between the detection of two photons, reveals important properties of the state to which they belonged, for example the fluctuations of the photon number. Since predictability is important for many applications, light sources emitting single photons are also characterized by the indistinguishability of consecutively emitted photons, or of two photons from separate emitters. In paper I we investigate blinking behaviour in quantum emitters, and its effect on the interference pattern and photon statistics with photons from two separate emitters. Blinking refers to an emitters transition into a non-emitting state, and subsequent transition back to an emitting state. We show that blinking can not be treated as linear loss, when measuring the fourth-order correlation function for two emitters in a Hong-Ou-Mandel setup. In general, a measurement of the fourth-order correlation function is robust to loss, which makes it a very practical tool. However, the relation between recorded coincidence counts and the correlation function is only direct in the limit of zero detection efficiency, and depends on the detection system. In paper II, we show that by adding a variable attenuation in the beam path, we can trace back to the ''true'' value of the correlation function at zero quantum efficiency. This method improves accuracy in correlation measurements by decreasing a systematic error at the expense of an increased statistical error, which is easier to handle, extending the use of coincidence methods to classical and non-classical multi-photon states.
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.