Single-photon source in micro- or nanoscale is the basic building block of on-chip quantum information and scalable quantum network. Enhanced spontaneous emission based on cavity quantum electrodynamics (CQED) is one of the key principles of realizing single-photon sources fabricated by micro- or nanophotonic cavities. Here we mainly review the spontaneous emission of single emitters in micro- or nanostructures, such as whispering gallery microcavities, photonic crystals, plasmon nanostructures, metamaterials, and their hybrids. The researches have enriched light-matter interaction as well as made great influence in single-photon source, photonic circuit, and on-chip quantum information.

Resonances in optical cavities are used to manipulate light propagation, enhance light-matter interaction, modulate quantum states, and so on. However, the index contrast between the traditional cavities and the host is generally not high, which to some extent limited their performances. By putting dielectric cavities into a host of zero-index materials, index contrast in principle can approach infinity. Here, we analytically deduced Mie resonance conditions at this extreme circumstance. Interestingly, we discovered a so-called resonance nesting effect, in which a set of cavities with different radii can possess the same type of resonance at the same wavelength. We also revealed previously unknown degeneracy between the 2l-TM (2l-TE) and 2l+1-TE (2l+1-TM) modes for " 0 ( 0) material, and the 2l-TM and 2l-TE for both " 0 and 0. Such extraordinary resonance nesting and degeneracy provide additional principles to manipulate cavity behaviors.

Federated learning (FL) becomes a promising machine learning paradigm. The impact of heterogeneous hardware specifications and dynamic states on the FL process has not yet been studied systematically. This paper presents the first large-scale study of this impact based on real-world data collected from 136k smartphones. We conducted extensive experiments on our proposed heterogeneity-aware FL platform namely FLASH, to systematically explore the performance of state-of-the-art FL algorithms and key FL configurations in heterogeneity-aware and -unaware settings, finding the following. (1) Heterogeneity causes accuracy to drop by up to 9.2% and convergence time to increase by 2.32×. (2) Heterogeneity negatively impacts popular aggregation algorithms, e.g., the accuracy variance reduction brought by q-FedAvg drops by 17.5%. (3) Heterogeneity does not worsen the accuracy loss caused by gradient-compression algorithms significantly, but it compromises the convergence time by up to 2.5×. ( 4) Heterogeneity hinders client-selection algorithms from selecting wanted clients, thus reducing effectiveness. e.g., the accuracy increase brought by the state-of-the-art client-selection algorithm drops by 73.9%. ( 5) Heterogeneity causes the optimal FL hyper-parameters to drift significantly. More specifically, the heterogeneity-unaware setting favors looser deadline and higher reporting fraction to achieve better training performance. ( 6) Heterogeneity results in non-trivial failed clients (more than 10%) and leads to participation bias (the top 30% of clients contribute 86% of computations). Our FLASH platform and data have been publicly open sourced.

In this paper, we exploit two-mode model to investigate the influence of quantum fluctuation on the self-trapping of the Bose-Einstein Condensate(BEC). We found that, even for the finite number of particles, the transition to self-trapping is still observed.In particular, because of the quantum fluctuation due to the finite number of particles, the critical phenomenon becomes fuzzily: the smaller the number of particles is, the more obviously fuzzily the critical phenomenon becomes. Quantum entanglement entropy is then introduced to depict the critical behavior near the transition point.