Intelligent code completion has become an essential research task to accelerate modern software development. To facilitate effective code completion for dynamically-typed programming languages, we apply neural language models by learning from large codebases, and develop a tailored attention mechanism for code completion. However, standard neural language models even with attention mechanism cannot correctly predict the outof-vocabulary (OoV) words that restrict the code completion performance. In this paper, inspired by the prevalence of locally repeated terms in program source code, and the recently proposed pointer copy mechanism, we propose a pointer mixture network for better predicting OoV words in code completion. Based on the context, the pointer mixture network learns to either generate a withinvocabulary word through an RNN component, or regenerate an OoV word from local context through a pointer component. Experiments on two benchmarked datasets demonstrate the effectiveness of our attention mechanism and pointer mixture network on the code completion task.

We demonstrate a quantum walk with time-dependent coin bias. With this technique we realize an experimental single-photon one-dimensional quantum walk with a linearly-ramped time-dependent coin flip operation and thereby demonstrate two periodic revivals of the walker distribution. In our beam-displacer interferometer, the walk corresponds to movement between discretely separated transverse modes of the field serving as lattice sites, and the time-dependent coin flip is effected by implementing a different angle between the optical axis of half-wave plate and the light propagation at each step. Each of the quantum-walk steps required to realize a revival comprises two sequential orthogonal coin-flip operators, with one coin having constant bias and the other coin having a time-dependent ramped coin bias, followed by a conditional translation of the walker. [4][5][6] plus the fundamental interest of being a natural quantized version of the ubiquitous random walk that appears in statistics, computer science, finance, physics, and chemistry. QW research has focused on evolution due to repeated applications of a time-independent unitary step operator U , but a QW with time-dependent unitary steps U (t), with discrete time t ∈ N := {0, 1, 2, . . . }, opens a much richer array of phenomena including localization and quasiperiodicity [7,8]. Here we demonstrate a time-dependent QW and use this technique to demonstrate a revival of the walker's position distribution.Rather than employing direct time-dependent control, we simulate time-dependent coin control by setting different coin parameters for different steps, which are effected in different locations along the longitudinal axis within our photonic beam-displacer interferometer (BDI) [9]. The quantum walker within the BDI is a single heralded photon produced by spontaneous parametric down conversion, and its walking degree of freedom is the set of discretely spaced transverse beam modes. The coin flip is effected by employing quarter-and half-wave plates.Our method for realizing the first time-dependent QW demonstrates the phenomenon of revivals and also opens the door to realizing a multitude of time-dependent QWs experimentally. Compared to prior work employing position-dependent control [10][11][12], our new technique decreases experimental complexity by relaxing the requirement of optical compensation. Our QW revival displays a different characteristic than typical QW properties such as ballistic spreading and localization of the walker distribution.The QW with a coin proceeds as a sequence of coin flips and then walker-coin entangling operations whereby the walker's position is displaced according to the coin state. We explain the QW now in full generality so the coin operator admits both spatial and temporal dependence. Spatially-dependent coin operations have dramatically demonstrated the realization of topological phases by QWs [4][5][6], but the time-dependent QW is, until now, only a theoretical construct and not yet explored experimentally.We employ a two...

No abstract

Our findings provides new information on baseline morphology of retinal microvasculature and its associated factors in school children, which will be useful for interpreting OCT-A metrics and for identifying and characterising pathological changes in retinal microvasculature.

IMPORTANCESecondhand smoking is a risk to adult ocular health, but its effect on children's ocular development is not known.OBJECTIVE To assess the association between choroidal thickness and secondhand smoking exposure in children. DESIGN, SETTING, AND PARTICIPANTSChildren aged 6 to 8 years were consecutively recruited from January 2016 to July 2017 from the population-based Hong Kong Children Eye Study at the Chinese University of Hong Kong Eye Centre. All participants underwent detailed ophthalmic investigations. Choroidal thickness was measured by swept-source optical coherence tomography, with built-in software that automatically segmented the choroid layer to analyze its terrain imagery. History of secondhand smoking was obtained from a questionnaire. Multiple linear regression analyses were performed to assess the correlation between choroidal thickness and secondhand exposure when controlling for confounding factors. Analysis began July 2018 and ended in April 2019. MAIN OUTCOMES AND MEASUREMENTSThe association between children's choroidal thickness and their exposure to secondhand smoking. RESULTSOf 1400 children, 941 (67.2%) had no exposure to secondhand smoking, and 459 (32.8%) had exposure to secondhand smoking. The mean (SD) age was 7.65 (1.09) years for children in the nonexposure group and 7.54 (1.11) years for children in the exposure group. After adjustment for age, sex, body mass index, axial length, and birth weight, exposure to secondhand smoking was associated with a thinner choroid by 8.3 μm in the central subfield, 7.2 μm in the inner inferior, 6.4 μm in the outer inferior, 6.4 μm in the inner temporal, and 7.3 μm in the outer temporal. Choroidal thinning with also associated with increased number of family smokers and increased quantity of secondhand smoking. An increase of 1 family smoker was associated with choroidal thinning by 7.86 μm in the central subfield, 4.51 μm in the outer superior, 6.23 μm in the inner inferior, 5.59 μm in the outer inferior, 6.06 μm in the inner nasal, and 6.55 μm in the outer nasal. An increase of exposure to 1 secondhand cigarette smoke per day was associated with choroidal thinning by 0.54 μm in the central subfield, 0.42 μm in the inner temporal, and 0.47 μm in the outer temporal.CONCLUSIONS AND RELEVANCE This investigation showed that exposure to secondhand smoking in children was associated with choroidal thinning along with a dose-dependent effect. These results support evidence regarding the potential hazards of secondhand smoking to children.

Uncertainty relations are the hallmarks of quantum physics and have been widely investigated since its original formulation. To understand and quantitatively capture the essence of preparation uncertainty in quantum interference, the uncertainty relations for unitary operators need to be investigated. Here, we report the first experimental investigation of the uncertainty relations for general unitary operators. In particular, we experimentally demonstrate the uncertainty relation for general unitary operators proved by Bagchi and Pati [ Phys. Rev. A94, 042104 (2016)], which places a non-trivial lower bound on the sum of uncertainties and removes the triviality problem faced by the product of the uncertainties. The experimental findings agree with the predictions of quantum theory and respect the new uncertainty relation.

Leggett-Garg inequalities are tests of macroscopic realism that can be violated by quantum mechanics. In this letter, we realise photonic Leggett-Garg tests on a three-level system and implement measurements that admit three distinct measurement outcomes, rather than the usual two. In this way we obtain violations of three- and four-time Leggett-Garg inequalities that are significantly in excess of those obtainable in standard Leggett-Garg tests. We also report violations the quantum-witness equality up to the maximum permitted for a three-outcome measurement. Our results highlight differences between spatial and temporal correlations in quantum mechanics.

No abstract