A novel temporal generative adversarial network for electrocardiography anomaly detection.

Cardiac abnormality detection from Electrocardiogram (ECG) signals is a common task for cardiologists. To facilitate efficient and objective detection, automated ECG classification by using deep learning based methods have been developed in recent years. Despite their impressive performance, these methods perform poorly when presented with cardiac abnormalities that are not well represented, or absent, in the training data. To this end, we propose a novel one-class classification based ECG anomaly detection generative adversarial network (GAN). Specifically, we embedded a Bi-directional Long-Short Term Memory (Bi-LSTM) layer into a GAN architecture and used a mini-batch discrimination training strategy in the discriminator to synthesis ECG signals. Our method generates samples to match the data distribution from normal signals of healthy group so that a generalised anomaly detector can be built reliably. The experimental results demonstrate our method outperforms several state-of-the-art semi-supervised learning based ECG anomaly detection algorithms and robustly detects the unknown anomaly class in the MIT-BIH arrhythmia database. Experiments show that our method achieves the accuracy of 95.5% and AUC of 95.9% which outperforms the most competitive baseline by 0.7% and 1.7% respectively. Our method may prove to be a helpful diagnostic method for helping cardiologists identify arrhythmias.

[Effects of deficit irrigation on seed production performance and water use efficiency of two native plant species in arid areas]. / äºç¼ºçŒæº‰å¯¹å¹²æ—±åŒºä¸¤ç§ä¹¡åœŸæ¤ç‰©ç§å­ç”Ÿäº§æ€§èƒ½åŠå ¶æ°´åˆ†åˆ©ç”¨æ•ˆçŽ‡çš„å½±å“.

Scientific irrigation is of great significance to plant seed production. With two excellent native plant species in desert steppe, Agropyron mongolicum and Lespedeza potaninii, as the objects, and full irrigation as the control, we explored the effects of deficit irrigation in different growth stages on the seed production and water use efficiency (WUE) of those two species. The results showed that, compared with the control, soil water content of both species decreased under deficit irrigation. The decrease of soil water content of A. mongolicum mainly occurred in the 0-60 cm soil layer, while there was no obvious stratification for the reduction of soil water content of L. potaninii. There were significant differences in the yield components of A. mongolica under deficit irrigation. The seed yield of deficit irrigation at the flowering stage was the highest. There were significant differences in the numbers of fertile tillers, florets and pods of L. potaninii among treatments, but no significant difference in seed yield. There were positive correlations between seed yield of A. mongolicum and the numbers of fertile tillers (r=0.776) and spikelets (r=0.717). The racemes of L. potaninii was significantly negatively correlated with the number of fertile tillers (r=-0.685), and positively correlated with the number of florets (r=0.412). Compared with full irrigation, water consumption of seed production of the two native plant species was reduced under deficit irrigation, but water use efficiency was improved, with the strongest improvement at the flowering stage of A. mongolicum (32.9%) and at the branching stage of L. potaninii (27.4%). Therefore, proper deficit irrigation could improve water use efficiency of both plant species. From the perspective of water use efficiency and seed yield, deficit irrigation could be used for artificial breeding of A. mongolicum and L. potaninii seeds in arid area, with the suitable growth stage for deficit being the flowering and the branching stages, respectively.

Stabilizing the active phase of iron-based Fischer-Tropsch catalysts for lower olefins: mechanism and strategy.

Fischer-Tropsch synthesis of lower olefins (FTO) is a classical yet modern topic of great significance in which the supported Fe-based nanoparticles are the most promising catalysts. The performance deterioration of catalysts is a big challenge due to the instability of the nanosized active phase of iron carbides. Herein, by in situ mass spectrometry, theoretical analysis, and atmospheric- and high-pressure experimental examinations, we revealed the Ostwald-ripening-like growth mechanism of the active phase of iron carbides in FTO, which involves the cyclic formation-decomposition of iron carbonyl intermediates to transport iron species from small particles to large ones. Accordingly, by suppressing the formation of iron carbonyl species with a high-N-content carbon support, the size and structure of the active phase were regulated and stabilized, and durable iron-based catalysts were conveniently obtained with the highest selectivity for lower olefins up to 54.1%. This study provides a practical strategy for exploring advanced FTO catalysts.

Electrocatalysis of S-doped carbon with weak polysulfide adsorption enhances lithium-sulfur battery performance.

Heteroatom-doped nanocarbons are beneficial for the performance improvement of lithium-sulfur batteries, and the reason is usually attributed to their strong adsorption to the soluble polysulfides. Herein, we found that, despite the weak polysulfide adsorption on hierarchical S-doped carbon nanocages (hSCNCs), the hSCNC-encapsulated sulfur cathode still exhibited better performance than the counterpart using undoped carbon nanocages, showing a high capacity of 579 mA h g-1 at 2 A g-1 after 400 cycles, and a high areal capacity of 4.7 mA h cm-2 with a high sulfur loading of 4.5 mg cm-2. The electrocatalysis-promoted mechanism of S-doped carbon was demonstrated, which facilitated polysulfide conversion and suppressed the polarization effect, thereby leading to superior performance.

Evolution of mixing width induced by general Rayleigh-Taylor instability.

Turbulent mixing induced by Rayleigh-Taylor (RT) instability occurs ubiquitously in many natural phenomena and engineering applications. As the simplest and primary descriptor of the mixing process, the evolution of mixing width of the mixing zone plays a notable role in the flows. The flows generally involve complex varying acceleration histories and widely varying density ratios, two dominant factors affecting the evolution of mixing width. However, no satisfactory theory for predicting the evolution has yet been established. Here a theory determining the evolution of mixing width in general RT flows is established to reproduce, first, all of the documented experiments conducted for diverse (i.e., constant, impulsive, oscillating, decreasing, increasing, and complex) acceleration histories and all density ratios. The theory is established in terms of the conservation principle, with special consideration given to the asymmetry of the volume-averaged density fields occurring in actual flows. The results reveal the sensitivity or insensitivity of the evolution of a mixing front of a neighboring light or heavy fluid to the degree of asymmetry and thus explain the distinct evolutions in two experiments with the same configurations.