No abstract

Coronavirus disease 2019 (COVID-19) has infected more than 1.3 million individuals all over the world and caused more than 106,000 deaths. One major hurdle in controlling the spreading of this disease is the inefficiency and shortage of medical tests. There have been increasing efforts on developing deep learning methods to diagnose COVID-19 based on CT scans. However, these works are difficult to reproduce and adopt since the CT data used in their studies are not publicly available. Besides, these works require a large number of CTs to train accurate diagnosis models, which are difficult to obtain. In this paper, we aim to address these two problems. We build a publicly-available dataset containing hundreds of CT scans positive for COVID-19 and develop sample-efficient deep learning methods that can achieve high diagnosis accuracy of COVID-19 from CT scans even when the number of training CT images are limited. Specifically, we propose an Self-Trans approach, which synergistically integrates contrastive self-supervised learning with transfer learning to learn powerful and unbiased feature representations for reducing the risk of overfitting. Extensive experiments demonstrate the superior performance of our proposed Self-Trans approach compared with several state-of-the-art baselines. Our approach achieves an F1 of 0.85 and an AUC of 0.94 in diagnosing COVID-19 from CT scans, even though the number of training CTs is just a few hundred.

No abstract

No abstract

We present radiation transfer simulations of evolutionary sequences of massive protostars forming from massive dense cores in environments of high mass surface densities, based on the Turbulent Core model (McKee & Tan 2003). The protostellar evolution is calculated with a multi-zone numerical model, with accretion rate regulated by feedback from an evolving disk-wind outflow cavity. Disk evolution is calculated assuming a fixed ratio of disk to protostellar mass, while core envelope evolution assumes inside-out collapse of the core of fixed outer radius. In this framework, an evolutionary track is determined by three environmental initial conditions: core mass M c , mass surface density of the ambient clump Σ cl , and ratio of the core's initial rotational to gravitational energy β c . Evolutionary sequences with various M c , Σ cl , β c are constructed. We find that in a fiducial model with M c = 60 M ⊙ , Σ cl = 1 g cm −2 and β c = 0.02, the final mass of the protostar reaches at least ∼ 26 M ⊙ , making the final star formation efficiency 0.43. For each of the evolutionary tracks, radiation transfer simulations are performed at selected stages, with temperature profiles, spectral energy distributions (SEDs), and multi-wavelength images produced. At a given stage, envelope temperature is depends strongly on Σ cl , with higher temperatures in a higher Σ cl core, but only weakly on M c . The SED and MIR images depend sensitively on the evolving outflow cavity, which gradually widens as the protostar grows. The fluxes at 100 µm increase dramatically, and the far-IR peaks move to shorter wavelengths. The influence of Σ cl and β c (which determines disk size) are discussed. We find that, despite scatter caused by different M c , Σ cl , β c , and inclinations, sources at a given evolutionary stage appear in similar regions of color-color diagrams, especially when using colors with fluxes at 70 µm, where scatter due to inclination is minimized, implying that such diagrams can be useful diagnostic tools of evolutionary stages of massive protostars. We discuss how intensity profiles along or perpendicular to the outflow axis are affected by environmental conditions and source evolution, and can thus act as additional diagnostics of the massive star formation process.

No abstract

The continuous-variable version of quantum key distribution (QKD) offers the advantages (over discrete-variable systems) of higher secret key rates in metropolitan areas, as well as the use of standard telecom components that can operate at room temperature. An important step in the real-world adoption of continuous-variable QKD is the deployment of field tests over commercial fibers. Here we report two different field tests of a continuousvariable QKD system through commercial fiber networks in Xi'an and Guangzhou over distances of 30.02 km (12.48 dB) and 49.85 km (11.62 dB), respectively. We achieve secure key rates two orders-of-magnitude higher than previous field test demonstrations by employing an efficient calibration model with one-time evaluation. This accomplishment is also realized by developing a fully automatic control system which stabilizes system noise, and by applying a rate-adaptive reconciliation method which maintains high reconciliation efficiency with high success probability in fluctuated environments. Our results pave the way to deploy continuousvariable QKD in metropolitan settings.

No abstract