ECG Heartbeat Classification Under Dataset Shift

Electrocardiogram(ECG)is widely used to detect arrhythmia.Atrial fibrillation,atrioventricular block,premature beats,etc.can all be diagnosed by ECG.When the distribution of training data and test data is inconsistent,the accuracy of the model will be affected.This phenomenon is called dataset shift.In the real-world heartbeat classification system,the heartbeat of the training set and test set often comes from patients of different ages and genders,so there are differences in the distribution of data sets.The main challenge in applying machine learning algorithms to clinical AI systems is dataset shift.Test-time adaptation(TTA)aims to adapt a pre-trained model from the source domain(SD)to the target domain(TD)without using any SD data or TD labels,thereby reducing model performance degradation due to domain differences.We propose a method based on multimodal image fusion and continual test-time adaptation(FCTA)for accurate and efficient heartbeat classification.First,the original ECG data is converted into a three-channel color image through a multimodal image fusion framework.The impact of class imbalance on network performance is overcome using a batch weight loss function,and then the pretrained source model is adapted to the TD using a continual test-time adaptation(CTA)method.Although our method is very simple,compared with other domain adaptation methods,it can significantly improve model performance on the test set and reduce the impact caused by the difference in domain distribution.

Improving Expectation Propagation with Lattice Reduction for Massive MIMO Detection

Efficient massive MIMO detection for practical deployment, which is with spatially correlated channel and high-order modulation, is a challenging topic for the fifth generation mobile communication(5 G). In this paper, we propose a lattice reduction aided expectation propagation(LRA-EP) algorithm for massive MIMO detection. LRA-EP applies expectation propagation in lattice reduced MIMO system to approach the distribution of lattice reduced constellation point by iterative refinement on its parameters(mean and covariance). The parameter refinement is based on the lattice reduced, well-conditioned MIMO channel. Numerical result shows that LRA-EP outperforms classic EP based MIMO detection(EPD) with 5～7 dB in terms of required signalto-noise ratio(SNR) for 1% packet error rate in spatially correlated channel for 256-QAM. We also show that LRA-EP has lower computation complexity than EPD.

Ergodic Rate Analysis on Applying Antenna Selection in D2D Communication Underlaying Cellular Networks

By reusing the spectrum of a cellular network, device-to-device(D2D) communications is known to greatly improve the spectral efficiency bypassing the base station(BS) of the cellular network. Antenna selection is the most cost efficient scheme for interference management, which is crucial to D2D systems. This paper investigates the achievable rate performance of the D2D communication underlaying the cellular network where a multiple-antenna base station with antenna selection scheme is deployed. We derive an exact closed-form expression of the ergodic achievable rate. Also, using Jensen's inequality, two pairs of upper and lower bounds of the rate are derived and we validate the tightness of the two sets of bounds. Based on the bounds obtained, we analyze the ergodic achievable rate in noise-limited scenario, interference-limited high SNR scenario and larger-scale antenna systems. Our analysis shows that the presence of D2D users could be counter-productive if the SNR at cellular UE is high. Further analysis shows that the relationship between the ergodic rate and the number of antennas it positive, but keeps decreasing as the antenna number increasing. These show the inefficiency of antenna selection in D2D interference management.

Mu-Net:Multi-Path Upsampling Convolution Network for Medical Image Segmentation

Medical image segmentation plays an important role in clinical diagnosis,quantitative analysis,and treatment process.Since 2015,U-Net-based approaches have been widely used formedical image segmentation.The purpose of the U-Net expansive path is to map low-resolution encoder feature maps to full input resolution feature maps.However,the consecutive deconvolution and convolutional operations in the expansive path lead to the loss of some high-level information.More high-level information can make the segmentationmore accurate.In this paper,we propose MU-Net,a novel,multi-path upsampling convolution network to retain more high-level information.The MU-Net mainly consists of three parts:contracting path,skip connection,and multi-expansive paths.The proposed MU-Net architecture is evaluated based on three different medical imaging datasets.Our experiments show that MU-Net improves the segmentation performance of U-Net-based methods on different datasets.At the same time,the computational efficiency is significantly improved by reducing the number of parameters by more than half.

Research progress and application of deep in-situ condition preserved coring and testing

With the depletion of shallow resources,the exploration of deep earth resources has become a global strategy.The study of the different patterns in the physical mechanical properties of rocks at different occurrence depths is the basis for exploring deep into the earth,with the core and premise being the acquisition and testing of deep in-situ core specimens.Based on the original idea of deep in-situ condition preserved coring(ICP-Coring)and testing,combined with theoretical modeling,numerical analysis,test platform development,indoor testing and engineering application,the principles and technologies of deep ICP-Coring are developed.This principle and technology consists of five parts:in-situ pressurepreserved coring(IPP-Coring),in-situ substance-preserved coring(ISP-Coring),in-situ temperaturepreserved coring(ITP-Coring),in-situ light-preserved coring(ILP-Coring),and in-situ moisturepreserved coring(IMP-Coring).The theory and technology of temperature and pressure reconstruction at different occurrence depths and in different environments are proposed,and prototype trial production was completed by following the principle of displacement and tests based on the in-situ reconstructed environment.The notable advances are as follows:(1)Deep in-situ coring system:A pressure-preserved controller with an ultimate bearing capacity greater than 140 MPa,highperformance(temperature-resistant,pressure-resistant,and low thermally conductive)temperaturepreserved materials,an active temperature control system,and high-barrier quality-preserved membrane materials were developed;a deep ICP-Coring capacity calibration platform was independently developed,a deep in-situ coring technology system was developed,and the acquisition of deep in-situ cores was realized.(2)In-situ storage displacement system:Following the dual-circuit hydraulic design idea,a single-drive source push-pull composite grabbing mechanism was designed;the design of the overall structure for the deep in-situ displacement storage system and ultrahigh pressure cabin structure was completed,which could realize docking the coring device and core displacement in the in-situ reconstructed environment.(3)Test analysis system:A noncontact acoustic-electric-magnetic test system was developed under the in-situ reconstructed environment,and the errors between the test results and traditional contact test results were mostly less than 10%;a detachable deep in-situ core true triaxial test system was developed,which could perform loading tests for deep in-situ cores.The relevant technological achievements were successfully applied to the exploration and development of deep resources,such as deep mines,deep-sea natural gas hydrates,and deep oil and gas.The research results provide technical and equipment support for the construction of a theoretical system for deep in-situ rock mechanics,the development of deep earth resources and energy,and the scientific exploration of different layers and occurrence depths(deep and ultradeep)of the Earth.

A Novel Feature Aggregation Approach for Image Retrieval Using Local and Global Features

The current deep convolution features based on retrievalmethods cannot fully use the characteristics of the salient image regions.Also,they cannot effectively suppress the background noises,so it is a challenging task to retrieve objects in cluttered scenarios.To solve the problem,we propose a new image retrieval method that employs a novel feature aggregation approach with an attention mechanism and utilizes a combination of local and global features.The method first extracts global and local features of the input image and then selects keypoints from local features by using the attention mechanism.After that,the feature aggregation mechanism aggregates the keypoints to a compact vector representation according to the scores evaluated by the attention mechanism.The core of the aggregation mechanism is to allow features with high scores to participate in residual operations of all cluster centers.Finally,we get the improved image representation by fusing aggregated feature descriptor and global feature of the input image.To effectively evaluate the proposedmethod,we have carried out a series of experiments on large-scale image datasets and compared them with other state-of-the-art methods.Experiments show that this method greatly improves the precision of image retrieval and computational efficiency.

Modulation Index Estimation of Frequency and Phase Modulated Signals

Modulation index estimation is important in the demodulation and recognition of Angle Modulation (AM) signals which include Frequency Modulation (FM) and Phase Modulation (PM) signals. In this paper, we firstly analyzed the AM signals with baseband modulation types, such as monotone, PSK, FSK, in the time and frequency domain. Then, we established a unified mathematical representation for the AM signals. Finally, we derived a blind estimation algorithm for the modulation index without using any prior knowledge. Simulation results verify the capabilities of the proposed algorithm.

Research and Development on an Embedded Digital Storage Oscillo-scope Based on PXA 270

This paper discusses the chief techniques and design principles of an embedded digital storage oscilloscope based on PXA 270, using FPGA as logic controller cell. The data acquisition partition consists of pre-process circuit, A/D converter, on-board memories, and control circuit integrated in FPGA. In the PB and EVC development environment, developers realized the part of em-bedded software. Actual test showed that the highest real-time sampling rate of the oscilloscope is up to lGHz/s, which has achieved the desired design requirements.

Study of Texture Segmentation and Classification for Grading Small Hepatocellular Carcinoma Based on CT Images

To grade Small Hepatocellular Car Cinoma(SHCC)using texture analysis of CT images,we retrospectively analysed 68 cases of Grade II(medium-differentiation)and 37 cases of Grades III and IV(high-differentiation).The grading scheme follows 4 stages:(1)training a Super Resolution Generative Adversarial Network(SRGAN)migration learning model on the Lung Nodule Analysis 2016 Dataset,and employing this model to reconstruct Super Resolution Images of the SHCC Dataset(SR-SHCC)images;(2)designing a texture clustering method based on Gray-Level Co-occurrence Matrix(GLCM)to segment tumour regions,which are Regions Of Interest(ROIs),from the original and SR-SHCC images,respectively;(3)extracting texture features on the ROIs;(4)performing statistical analysis and classifications.The segmentation achieved accuracies of 0.9049 and 0.8590 in the original SHCC images and the SR-SHCC images,respectively.The classification achived an accuracy of 0.838 and an Area Under the ROC Curve(AUC)of 0.84.The grading scheme can effectively reduce poor impacts on the texture analysis of SHCC ROIs.It may play a guiding role for physicians in early diagnoses of medium-differentiation and high-differentiation in SHCC.

JAG1 enhances angiogenesis in triple-negative breast cancer through promoting the secretion of exosomal lncRNA MALAT1

Despite significant improvements in five-year survival rates due to early diagnosis and combination therapy, triple-negative breast cancer (TNBC) treatment remains a major challenge. Finding new and effective targets for diagnosis and drug therapy is urgent for TNBC patients. Jagged-1 (JAG1), one of the canonical ligands of the Notch signaling pathway, is involved in vascular budding and is a poor prognostic factor of TNBC. In this study, combined with quantitative real-time PCR, database analysis, animal experiments, and other means, JAG1 was confirmed to be related to the poor prognosis of TNBC patients. JAG1 was highly expressed in MDA-MB-231 Bone (231B) cells, with stronger invasion and metastasis ability than MDA-MB-231 (231) cells. Treatment of human vascular endothelial cells (HUVEC) with TNBC conditioned medium showed that TNBC JAG1 promoted the angiogenesis of HUVEC. Next, we detected the exosomes extracted from TNBC conditioned medium and found that JAG1 promoted the exosome secretion from 231 cells via ALIX-RAB11A/RAB35. In addition, we also found that the exosomes from JAG1 overexpressed TNBC cells contained more long non-coding RNA (lncRNA) MALAT1 , and MALAT1 promoted angiogenesis of HUVEC by targeting miR-140-5p . Finally, the angiogenesis-promoting effect of JAG1 in TNBC was further investigated by matrix gel assay. In conclusion, we reveal that JAG1 has a pro-invasion effect on TNBC and is involved in microenvironment angiogenesis by promoting exosome secretion and the MALAT1-miR-140-5p-JAG1/VEGFA pathway.

Analytical method of nonlinear coupled constitutive relations for rarefied non-equilibrium flows

It is well known that Navier-Stokes equations are not valid for those high-Knudsen and high-Mach flows, in which the local thermodynamically non-equilibrium effects are dominant. To extend the non-equilibrium describing the ability of macroscopic equations, Nonlinear Coupled Constitutive Relation(NCCR) model was developed from Eu’s generalized hydrodynamic equations to substitute linear Newton’s law of viscosity and Fourier’s law of heat conduction in conservation laws. In the NCCR model, how to solve the decomposed constitutive equations with reasonable computational cost is a key ingredient of this scheme. In this paper, an analytic method is proposed firstly. Compared to the iterative procedure in the conventional NCCR model, the analytic method not only obtains exact roots of the decomposed constitutive polynomials, but also preserves the nonlinear constitutive relations in the original framework of NCCR methods. Numerical tests to assess the efficiency and accuracy of the proposed method are conducted for argon shock structures, Couette flows, two-dimensional hypersonic flows over a cylinder and threedimensional supersonic flows over a three-dimensional sphere. These superior advantages of the current method are expected to render itself a powerful tool for simulating the hypersonic rarefied flows and microscale flows of high Knudsen number for engineering applications.

Bioresponsive nanogated ensemble based on structure-switchable aptamer directed assembly and disassembly of gold nanoparticles from mesoporous silica supports

By taking advantage of recent advances in aptamer biology and nanotechnology, we developed a general approach for the design and fabrication of bioresponsive controlled delivery systems. It utilized the structure-switchable aptamer directed assembly and disassembly of gold nanoparticles from mesoporous silica supports, which enables the control of cargo release from the inside of the mesoporous nanoparticles specifically in the presence of target molecule.

Improved Calculation of Turbulence Parameters Based on Six Tropical Cyclone Cases:Implication to Wind Turbine Design in Typhoon-Prone Areas

In view of the absence or insufficiency of tropical cyclone(TC) turbulence parameters in current design standards of wind turbines, in this paper, TC turbulence parameter models with roughness length involved are developed based on six landfall TCs observed from meteorological towers located on various underlying surfaces, so as to provide references for the wind turbine design under TC conditions. Firstly, the roughness length values are examined in order to reduce the effect on turbulence parameters of the various underlying surfaces. On this basis, the reference turbulence intensity is normalized by the roughness length. The related turbulence parameters are parameterized, including the turbulence standard deviation and the turbulence spectrum;and the turbulence parameters available under TC conditions for turbulence turbine design are presented finally. Comparisons of the wind parameter models presented in this paper with those used in current turbine design standards suggest that the former can represent TC characteristics more accurately. In order to withstand TCs, we suggest that the turbulence parameter models recommended in this paper be included in future wind turbine design standards under TC conditions.

超大跨度异型钢连桥动态成型施工关键技术研究与应用

为重塑西安成为“一带一路”文化核心区,扩大文化交流在西安市内的开展面和贯入度,西安市政府建立“三中心”重点配套公共服务类建设项目———“一带一路”文化交流中心系列公建项目。“一带一路”文化交流中心系列公建项目北地块为400 m超长非对称双塔连体结构,其中150 m空中连桥是国内最大跨度的钢连桥单体。钢连桥重约42000 kN,与两端塔楼焊接连接,桥身主体由片式桁架、弧形三角桁架及其之间的水平梁撑组成。在连桥一端下部加设有36 m长、4500 kN重的下加腋结构。连桥投影下方为单、双层地下室的多标高顶板,施工条件复杂。现阶段提升施工的研究,多停留在跨度小、重量轻的连体结构整体提升的理论及实践分析上,而对于包含有上述非常规自身结构特点及各类外部复杂环境的超长超重连桥施工,未有成套的关键技术研究和应用成果。经过对钢连桥施工重、难点的全面分析,结合行业内提升施工的现有案例,创新研发了一套超大跨度异型钢连桥动态成型施工关键技术,包含超大跨度异型钢连桥动态成型线形控制技术、超大跨度异型钢连桥结构内力控制与应力监测技术、复杂工况下大体量钢连桥拼装施工危险源管控技术。异型钢连桥在动态成型过程中,加腋结构分步安装使得结构重心持续变动,对于结构成型线形控制提出了更大挑战。150 m大跨度连桥由多节段厚板钢材构件拼装成整体,经分阶段提升动态成型。成型过程中,内力不断重分布,因此施工过程的内力控制与实时监测至关重要。此外,连桥体量大,多数构件单重达300 kN,其拼装施工对于大型起重设备的需求与其投影下方地下室顶板相对较弱的承载能力之间的矛盾,对施工安全危险源管控技术提出了更高要求。该技术提出了非对称预拱技术、优化了焊接顺序优化、应用循环预拼装及虚拟预拼装,解决了非对称超长连桥线形控制控制难题;应用全周期仿真模拟,研发了“多吊点分布、主被动力结合”的提升体系,开发了“差值法”应力判别标准,解决了施工过程中结构内力与应力监测难题;提出了“钢支撑+后浇承台”楼板加固,研发了多标高拼装平台体系,确保施工安全性。

Tighter bounds of multihop relayed communications in Nakagami-m fading

Closed-form bounds for the end-to-end performance of multihop communications with non-regenerative relays over Nakagami-m fading channels are investigated.Upper and lower bounds of the end-to-end signal-to-noise ratio(SNR)are first developed by using the monotonicity.Then,the probability density functions(PDFs),the cumulative distribution functions,and the momentgenerating functions(MGFs)of the bounds are derived.Using these results,the bounds for the outage and average bit error probability(ABEP)are obtained.Numerical and simulation results are executed to validate the tightness of the proposed bounds.