Speaker
Description
We present progress on exploiting photon-number statistics to characterize ensembles of distinguishable light emitters beyond conventional resolution limits. The photon-number distribution of light collected from an ensemble can carry statistical signatures that reflect the properties of its constituents, even when relevant degrees of freedom such as time, frequency, or spatial mode cannot directly be resolved. A well-known example is the use of the second-order correlation function, $g^{(2)}$, to estimate the number of independent single-photon emitters within an ensemble.
Building on this idea, we develop statistical models of photon emission to identify and quantify individual emitters from photon-number measurements, and we analyze the information-theoretic bounds governing this inference problem. We show that emitter identification remains feasible even when the emitters overlap within the detector response and are not individually resolvable.
To demonstrate these ideas experimentally, we generate photons in ultrafast time bins separated by only a few picoseconds using group-velocity delays introduced by birefringent materials. The time bins follow thermal photon statistics produced by a spontaneous parametric down-conversion process. We detect the resulting light with photon-number-resolving transition-edge sensors capable of resolving up to ten photons within a single optical pulse. Although the detector is far too slow to resolve the individual time bins, we show that our statistical inference techniques can accurately retrieve the mean photon numbers of each bin from the measured photon-number distributions.
Finally, to extend this framework beyond analytically tractable models, we employ normalizing-flow neural networks capable of learning arbitrary probability distributions. This approach allows us to incorporate realistic experimental effects such as optical loss, detector demultiplexing, and noise, enabling the analysis of increasingly complex emitter ensembles, including both thermal and single-photon sources.
| Keyword-1 | Quantum Sensing |
|---|---|
| Keyword-2 | Photon Statistics |