An Electrocardiogram Signal Classification Algorithm Based on Improved Deep Residual Shrinkage Networks北大核心CSCD

心电信号分类是医疗保健领域的重要研究内容.针对大多数方法不能很好地降低样本数量少的类别漏诊率,以及降低预处理操作的复杂性问题,提出了一种基于改进深度残差收缩网络(IDRSN)的心电信号分类算法(即DRSL算法).首先,使用合成少数类过采样技术(SMOTE)扩充数量少的类别样本,从而解决了类不平衡问题;其次,利用改进深度残差收缩网络提取空间特征,其残差模块可以避免网络层加深造成的过拟合,压缩激励和软阈值化子网络可以提取重要局部特征并自动去除噪声;然后,通过长短期记忆网络(LSTM)提取时间特征;最后,利用全连接网络输出分类结果.在MIT-BIH心律失常数据集上的实验结果表明,该算法的分类性能优于IDRSN、DRSN、GAN+2DCNN、CNN+LSTM_ATTENTION、SE-CNN-LSTM分类算法.     

深海低声速沉积层简正波用于海底参数反演*

现有深海地声参数反演研究通常基于全波场理论,存在计算量大、多值性以及需要准确的水文环境信息等问题。针对这一问题,提出一种基于简正波频散特征的深海低声速沉积层海底参数反演方法。在南海北部大陆坡海域的一次实验中,坐底式水听器(深度约1740 m)接收的沉底弹信号中观察到一种低频成分先到而高频成分后至的到达结构,由被限制在低声速沉积层的简正波组成。通过匹配沉积层简正波60-220 Hz频段内的到达时间差对实验海域的低声速海底参数进行反演,得到沉积层厚度为16.4 m,沉积层声速为1450.5 m/s,与底质采样数据比较吻合,且具有较高的可信度,验证了所提方法的有效性。由于沉积层密度对简正波频散特征的敏感性较差,其反演结果可信度不是很高,需要进一步研究。     

基于深度学习的材料超声回波衰减预测方法

材料超声回波衰减是评价材料均匀一致性的常用方法, 针对具有复杂结构的航空发动机盘件难以进行材料底面超声回波衰减评价的问题, 本文提出了利用超声背散射波信号直接预测底面回波衰减的方法。采用10MHz聚焦探头进行超声背散射波数据的采集, 利用深度学习技术构建和训练模型,建立了基于深度学习的材料底面回波衰减预测方法, 同时讨论了采用不同信号形式的超声波信号分类识别模型的准确率差异。研究发现：基于深度学习技术可实现通过超声背散射波预测材料的底面回波衰减, 预测结果和实际底面回波衰减试验结果具有良好的一致性。     

Quantum Bayesian computation

Quantum Bayesian computation is an emerging field that levers the computational gains available from quantum computers. They promise to provide an exponential speed-up in Bayesian computation. Our article adds to the literature in three ways. First, we describe how quantum von Neumann measurement provides quantum versions of popular machine learning algorithms such as Markov chain Monte Carlo and deep learning that are fundamental to Bayesian learning. Second, we describe quantum data encoding methods needed to implement quantum machine learning including the counterparts to traditional feature extraction and kernel embeddings methods. Third, we show how quantum algorithms naturally calculate Bayesian quantities of interest such as posterior distributions and marginal likelihoods. Our goal then is to show how quantum algorithms solve statistical machine learning problems. On the theoretical side, we provide quantum versions of high dimensional regression, Gaussian processes and stochastic gradient descent. On the empirical side, we apply a quantum FFT algorithm to Chicago house price data. Finally, we conclude with directions for future research.     

基于PCAU-Net的快速多通道磁共振成像方法

多通道磁共振成像方法采用多个接收线圈同时欠采样k空间以加快成像速度,并基于后处理算法重建图像,但在较高加速因子时,其图像重建质量仍然较差．本文提出了一种基于PCAU-Net的快速多通道磁共振成像方法,将单通道实数U型卷积神经网络拓展到多通道复数卷积神经网络,设计了一种结构不对称的U型网络结构,通过在解码部分减小网络规模以降低模型的复杂度．PCAU-Net网络在跳跃连接前增加了1×1卷积,以实现跨通道信息交互．输入和输出之间利用残差连接为误差的反向传播提供捷径．实验结果表明,使用规则和随机采样模板,在不同加速因子时,相比常规的GRAPPA重建算法和SPIRiT重建方法,本文提出的PCAU-Net方法可高质量重建出磁共振复数图像,并且相比于PCU-Net方法,PCAU-Net减少了模型参数、缩短了训练时间．     

基于DCGANs的半片光伏组件电致发光图像增强技术

针对半片光伏组件电致发光（electroluminescence,EL）缺陷自动识别过程中训练用样本不足导致模型过拟合的问题,采用深度卷积生成对抗网络（deep convolutional generative adversarial networks,DCGANs）生成可控制属性的半片光伏组件EL图像,再采用多尺度结构相似性（multiscale structural similarity,MS-SSIM）指标对生成的EL图像与拍摄的EL图像之间的相似程度进行了评估。评估结果得到,使用DCGANs生成的所有类型半片光伏组件的EL图像与拍摄的EL图像的MS-SSIM指标都大于0.55,大部分的MS-SSIM值在0.7附近。在分类模型的训练过程中,测试集准确率随着训练集中生成图像数量的增加而升高,当生成图像数量达到6 000张时,测试集准确率达到97.92%。实验结果表明,采用DCGANs能够生成高质量且可控制属性的半片光伏组件EL图像,较好地解决因缺少训练样本而导致的模型过拟合问题。     

Search for charm-quark production via dimuons in neutrino telescopes

Dimuon events induced by charm-quark productions from neutrino deep inelastic scattering (DIS) processes have been studied in traditional DIS experiments for decades. The recent progress in neutrino telescopes makes it possible to search for such dimuon events at energies far beyond the laboratory scale. In this study, we construct a simulation framework to calculate yields and distributions of dimuon signals in an IceCube-like km³ scale neutrino telescope. Owing to the experimental limitation in the resolution of double-track lateral distance, only dimuons produced outside the detector volume are considered. Detailed information about simulation results for a 10-year exposure is presented. As an earlier paper[Physical Review D 105, 093005 (2022)] and ours report on a similar situation, we use that paper as a baseline to conduct comparisons. We then estimate the impacts of different calculation methods of muon energy losses. Finally, we study the experimental potential of dimuon searches under the hypothesis of single-muon background only. Our results based on a simplified double-track reconstruction indicate a moderate sensitivity, especially with the ORCA configuration. Further developments on both the reconstruction algorithm and possible detector designs are thus required and are under investigation.     

剪切弯曲下短深梁位移数值计算精度的研究

分析了短深梁位移的基本解，评述了有限元解法中关于高次分布力简化为节点力的问题。使用虚功原理和位移线性插值推导了满足二次分布力作用的精确节点力分配公式。通过矩形截面悬臂梁自由终端受集中载荷的算例，阐明了提高短深梁位移计算精度的途径。     

Electrodeposition and characterization of Zn–Sn alloy coatings from a deep eutectic solvent based on choline chloride for corrosion protection

With the aim of obtaining high corrosion resistant Zn–Sn alloy coatings from an ionic liquid, the effects of electrodeposition potential and electrolyte composition on the electrodeposition behavior, film composition, morphology and corrosion performance were investigated. Cyclic voltammograms indicate that Zn and Sn were co‐deposited at distinct reduction potentials as pure Zn and Sn elements. In addition, the phase composition analysis also showed that the obtained Zn–Sn alloy deposits (8 wt.%–45 wt.% Zn) consist of a two‐phase mechanical mixture of small aggregates of Zn and Sn metals. The Zn content of the alloy significantly increases as the electrodeposition potential and electrolyte Zn (II)/Sn (II) ratio increase. The corrosion performance study of the obtained Zn–Sn coatings showed that they have a passivation behavior and their corrosion resistance increases as the alloy‐Sn content increases. To improve their morphological properties, ethylene diamine tetraacetic acid additive was introduced into the electrolyte and greatly improved the morphology and corrosion resistance of the deposits. For the first time, it was shown that high corrosion resistance Zn–Sn coatings can be obtained from ionic liquids. Copyright © 2014 John Wiley & Sons, Ltd.     

Magnetic graphene oxide modified with choline chloride-based deep eutectic solvent for the solid-phase extraction of protein

Four kinds of green deep eutectic solvents (DESs) based on choline chloride (ChCl) have been synthesized and coated on the surface of magnetic graphene oxide (Fe₃O₄@GO) to form Fe₃O₄@GO-DES for the magnetic solid-phase extraction of protein. X-ray diffraction (XRD), vibrating sample magnetometer (VSM), Fourier transform infrared spectrometry (FTIR), field emission scanning electron microscopy (FESEM) and thermal gravimetric analysis (TGA) were employed to characterize Fe₃O₄@GO-DES, and the results indicated the successful preparation of Fe₃O₄@GO-DES. The UV–vis spectrophotometer was used to measure the concentration of protein after extraction. Single factor experiments proved that the extraction amount was influenced by the types of DESs, solution temperature, solution ionic strength, extraction time, protein concentration and the amount of Fe₃O₄@GO-DES. Comparison of Fe₃O₄@GO and Fe₃O₄@GO-DES was carried out by extracting bovine serum albumin, ovalbumin, bovine hemoglobin and lysozyme. The experimental results showed that the proposed Fe₃O₄@GO-DES performs better than Fe₃O₄@GO in the extraction of acidic protein. Desorption of protein was carried out by eluting the solid extractant with 0.005 mol L⁻¹ Na₂HPO₄ contained 1 mol L⁻¹ NaCl. The obtained elution efficiency was about 90.9%. Attributed to the convenient magnetic separation, the solid extractant could be easily recycled.