InGaZnO-based photoelectric synaptic devices for neuromorphic computing

Photoelectric synaptic devices could emulate synaptic behaviors utilizing photoelectric effects and offer promising prospects with their high-speed operation and low crosstalk. In this study, we introduced a novel InGaZnO-based photoelectric memristor. Under both electrical and optical stimulation, the device successfully emulated synaptic characteristics including excitatory postsynaptic current (EPSC), paired-pulse facilitation (PPF), long-term potentiation (LTP), and long-term depression (LTD). Furthermore, we demonstrated the practical application of our synaptic devices through the recognition of handwritten digits. The devices have successfully shown their ability to modulate synaptic weights effectively through light pulse stimulation, resulting in a recognition accuracy of up to 93.4%. The results illustrated the potential of IGZO-based memristors in neuromorphic computing, particularly their ability to simulate synaptic functionalities and contribute to image recognition tasks.

基于时空Transformer-encoder的跨社交网络用户匹配方法

针对目前基于签到时空数据的跨社交网络用户匹配方法未充分利用时空信息之间的耦合关系,导致时空数据特征提取困难,匹配准确率下降的问题,提出了一种基于时空Transformer-encoder的跨社交网络用户匹配方法。该方法通过网格映射将签到时空信息转换为序列数据,生成签到序列;利用序列嵌入层将离散的签到序列映射到连续高维空间;然后借助多头注意力机制和卷积神经网络提取高维签到特征,并利用卷积神经网络实现优化多头注意力模块权重变换和特征融合;最后利用前馈神经网络实现分类,输出用户匹配得分。在两组真实社交网络用户数据集上进行大量用户匹配实验,与现有方法相比,准确率提升了0.40~10.53百分点,F_(1)值提升了0.43~9.5百分点。这验证了所提方法能够有效提取用户签到耦合特征,并提高用户匹配的性能。

A Lightweight Network with Dual Encoder and Cross Feature Fusion for Cement Pavement Crack Detection

Automatic crack detection of cement pavement chiefly benefits from the rapid development of deep learning,with convolutional neural networks(CNN)playing an important role in this field.However,as the performance of crack detection in cement pavement improves,the depth and width of the network structure are significantly increased,which necessitates more computing power and storage space.This limitation hampers the practical implementation of crack detection models on various platforms,particularly portable devices like small mobile devices.To solve these problems,we propose a dual-encoder-based network architecture that focuses on extracting more comprehensive fracture feature information and combines cross-fusion modules and coordinated attention mechanisms formore efficient feature fusion.Firstly,we use small channel convolution to construct shallow feature extractionmodule(SFEM)to extract low-level feature information of cracks in cement pavement images,in order to obtainmore information about cracks in the shallowfeatures of images.In addition,we construct large kernel atrous convolution(LKAC)to enhance crack information,which incorporates coordination attention mechanism for non-crack information filtering,and large kernel atrous convolution with different cores,using different receptive fields to extract more detailed edge and context information.Finally,the three-stage feature map outputs from the shallow feature extraction module is cross-fused with the two-stage feature map outputs from the large kernel atrous convolution module,and the shallow feature and detailed edge feature are fully fused to obtain the final crack prediction map.We evaluate our method on three public crack datasets:DeepCrack,CFD,and Crack500.Experimental results on theDeepCrack dataset demonstrate the effectiveness of our proposed method compared to state-of-the-art crack detection methods,which achieves Precision(P)87.2%,Recall(R)87.7%,and F-score(F1)87.4%.Thanks to our lightweight crack detectionmodel,the parameter count of the model in real-world detection scenarios has been significantly reduced to less than 2M.This advancement also facilitates technical support for portable scene detection.

Remaining Useful Life Prediction of Rail Based on Improved Pulse Separable Convolution Enhanced Transformer Encoder

In order to prevent possible casualties and economic loss, it is critical to accurate prediction of the Remaining Useful Life (RUL) in rail prognostics health management. However, the traditional neural networks is difficult to capture the long-term dependency relationship of the time series in the modeling of the long time series of rail damage, due to the coupling relationship of multi-channel data from multiple sensors. Here, in this paper, a novel RUL prediction model with an enhanced pulse separable convolution is used to solve this issue. Firstly, a coding module based on the improved pulse separable convolutional network is established to effectively model the relationship between the data. To enhance the network, an alternate gradient back propagation method is implemented. And an efficient channel attention (ECA) mechanism is developed for better emphasizing the useful pulse characteristics. Secondly, an optimized Transformer encoder was designed to serve as the backbone of the model. It has the ability to efficiently understand relationship between the data itself and each other at each time step of long time series with a full life cycle. More importantly, the Transformer encoder is improved by integrating pulse maximum pooling to retain more pulse timing characteristics. Finally, based on the characteristics of the front layer, the final predicted RUL value was provided and served as the end-to-end solution. The empirical findings validate the efficacy of the suggested approach in forecasting the rail RUL, surpassing various existing data-driven prognostication techniques. Meanwhile, the proposed method also shows good generalization performance on PHM2012 bearing data set.

Rational design of photofunctional dyes BODIPYs/aza-BODIPYs and applications for photocatalysis, photoelectric conversion and thermochromic materials

4,4-Difluoro-4-bora-3a,4a-diaza-sindacene (BODIPY) is a sort of photofunctional dye which possesses advantages including strong light-capturing property, high photon-resistance, etc. Meso-N substituted aza-BODIPY is a crucial derivative of BODIPY scaffold that has the favorable optical properties and a significant spectral redshift. The photophysical properties can be tuned by molecular design, and the attenuation path of the excited state energy release of absorbed light energy can be well controlled via structural modifications, enabling tailored application. It has been extensively employed in life medicine fields including fluorescence imaging diagnosis, photodynamic therapy photosensitizer and photothermal therapy reagent and so forth. Extensive research and review have been performed in these areas. However, BODIPYs/aza-BODIPYs have a significant role in energy, catalysis, optoelectronics, photo-responsive materials and other fields. Nevertheless, there are relatively few studies and reviews in these fields on the modification and application based on BODIPY/aza-BODIPY scaffold. Herein, in this review we summarized the application of BODIPY/aza-BODIPY in the aforementioned fields, with the molecular regulation of dye as the foundation and the utilization in the above fields as the objective, in the intention of providing inspiration for the exploration of innovative BODIPY/aza-BODIPY research in the field of light resource conversion and functional materials.

Virus-Encoded MicroRNAs Reveal How Ranavirus Interacts with Amphibian Immune Defense

Ranaviruses are harmful viruses that infect amphibians, fish, and reptiles, and have caused particularly devastating declines in amphibian populations. One particular type of ranavirus, called Frog Virus 3 (FV3), has been extensively studied due to its prevalence and impact on amphibians. Previous research has primarily focused on the virus’s genes, but little attention has been given to the non-coding regions of its genome. This article reviews recent studies that reveal the ability of ranaviruses, including FV3, to encode microRNA (miRNA), a type of regulatory RNA. These viral miRNAs play a crucial role in suppressing frog immune genes, modulating the virus-host interaction, and promoting viral infection. Understanding how ranaviruses use miRNAs to control disease progression is essential for addressing the health threat they pose to wildlife and ecosystems.

ENCODE计划和功能基因组研究

人类基因组计划完成以来,科学家们一直在努力阐释基因组信息所代表的生物学意义。自2003年开始,美国国家人类基因组研究所(National Human Genome Research Institute,NHGRI)投资近3亿美元启动"DNA元件百科全书(Encyclopedia of DNA Elements,ENCODE)"计划,集结了来自美国、中国、英国、日本、西班牙和新加坡等国家的32个实验室的440余名科学家,共同鉴定并分析人类基因组中所有的功能调控元件。高通量测序技术等实验手段的发展和生物信息学技术的不断完善使得ENCODE计划取得了丰硕的成果:确定了甲基化和组蛋白修饰等表观修饰区域及其对染色质结构的作用,进而确定染色质结构的改变影响基因表达;确定了转录因子及其结合位点的信息,并构建了转录因子调控网络;进一步修订更新了假基因和非编码RNA数据库;并确定了调控序列的单核苷酸多态性(Single nucleotide polymorphism,SNP)并与疾病相关联。这些发现一方面有助于系统解析基因和基因组信息、调控元件的调控作用以及非编码区转录调控等分子机制;同时也将为转化医学等生命科学研究领域提供丰富的数据来源。文章综述了高通量测序技术等实验手段的发展和生物信息学技术的不断完善对ENCODE计划的贡献、表观遗传学研究与ENCODE计划的关联性、ENCODE计划的主要科学成果等,同时展望了ENCODE计划对基础医学、临床医学和转化医学等生命科学研究领域的巨大推动作用。

ASP源代码加密程序Script Encoder算法研究

微软为了保护脚本代码的安全性 ,以 COM组件的形式提供了一种对脚本代码进行编码加密的技术 .但其安全性到底如何呢 ?本文将通过“反编译”的方式对其加密。

一种融合AutoEncoder与CNN的混合算法用于图像特征提取

深度学习方法在图像的特征提取方面具有优势。针对传统特征提取方法需要先验知识的不足,提出一种自动编码器(Auto Encoder)与卷积神经网络(convolutional neural network,CNN)相结合的深度学习特征提取方法。该方法给Auto Encoder加入快速稀疏性控制,据此对图像训练出基本构件,并初始化CNN的卷积核;同时,给CNN加入了滤波机制,使输出特征保持稀疏性。实验结果表明,在Minist手写数字库和Yale人脸库的识别效果上,提出的特征提取方法均取得了较好的结果,实验进一步通过交叉验证T检验指出,引入滤波机制的特征提取模型优于没有采用滤波机制的模型。

基于Windows Media Encoder实现流媒体同步控制

基于Windows Media技术的流媒体同步控制需要ASF的支持。现有方法要求直接修改ASF文件的头对象而不易编程实现。通过对比研究,提出了一种基于Windows Media Encoder编码器后处理过程间接修改ASF文件的方法,并探讨了该方法的基本原理。实际系统中的成功应用证明了该方法能有效降低编程难度并易于集成,对实现流媒体同步控制具有一定的参考价值。

基于ENCODER_ATT机制的远程监督关系抽取

在信息抽取中,关系抽取是一项准确识别自然语言中实体间关系的关键技术。针对关系抽取模型中容易丢失关键语义特征问题及远程监督的基本假设容易引入噪声数据的问题,本文提出一种基于远程监督的ENCODER_ATT关系抽取模型。基于循环神经网络构造的ENCODER模型在以词级别进行特征记忆提取,并在句子层面进行语义特征信息整合,保证不遗失关键语义特征的同时去除冗余特征。然后在句子层面引入了注意力机制来降低噪声数据对实验结果的影响。在真实的数据集上进行实验,并绘制准确率-召回率曲线,实验结果表明ENCODER_ATT模型对比同类型的关系抽取方法有明显的提升。

基于encoder-decoder框架的城镇污水厂出水水质预测

由于污水厂的出水水质指标繁多、污水处理过程中反应复杂、时序非线性程度高,基于机理模型的预测方法无法取得理想效果。针对此问题,提出基于深度学习的污水厂出水水质预测方法,并以吉林省某污水厂监测水质为来源数据,利用多种结合encoder-decoder结构的神经网络预测水质。结果显示,所提结构对LSTM和GRU网络预测能力都有一定提升,对长期预测能力提升更加显著,ED-GRU模型效果最佳,短期预测中的4个出水水质指标均方根误差(RMSE)为0.7551、0.2197、0.0734、0.3146,拟合优度(R2)为0.9013、0.9332、0.9167、0.9532,可以预测出水质局部变化,而长期预测中的4个指标RMSE为1.7204、1.7689、0.4478、0.8316,R2为0.4849、0.5507、0.4502、0.7595,可以预测出水质变化趋势,与顺序结构相比,短期预测RMSE降低10%以上,R2增加2%以上,长期预测RMSE降低25%以上,R2增加15%以上。研究结果表明,基于encoder-decoder结构的神经网络可以对污水厂出水水质进行准确预测,为污水处理工艺改进提供技术支撑。

基于Auto Encoder的智能监控指纹识别系统

针对目前已有的嵌入式指纹识别系统往往采用手工提取,不能自动学习并提取识别所需的特征及识别正确率仍然不高的缺点,提出一种基于自动编码器(Auto Encoder)和LSSVM的指纹识别系统。首先,提出采用FPS200作为指纹传感器采集指纹数据,然后将采集的数据经过滤波和二值化等预处理,通过比较差异算法获得Auto Encoder中的权值和偏置等参数,从而得到训练好的Auto Encoder用于指纹图像特征提取。最后,将自动提取的特征进行训练和分类,将投票最多的分类作为指纹识别的结果。通过测试表明,系统能较精确地实现指纹识别,具有收敛速度快、正确识别率高和匹配时间短的优点。

在Flash中基于Adobe Media Encoder组件的视频导入与应用方法

在flash cs6的默认情况下,Flash cs6只支持flv和f4v格式的视频.如果不是这种格式的视频,我们可以使用Flash cs6自带的视频转换组件Adobe Media Encoder将其他视频格式转换成FLV和F4V格式.本文主要讲解如何使用flash自带的Adobe Media Encoder组件进行视频文件的转换,导入和使用.

基于时空特征融合的Encoder-Decoder多步4D短期航迹预测

航迹预测在确保空中交通安全、高效运行中扮演着至关重要的角色。所预测的航迹信息是航迹优化、冲突告警等决策工具的输入,而预测准确性取决于模型对航迹序列特征的提取能力。航迹序列数据是具有丰富时空特征的多维时间序列,其中每个变量都呈现出长短期的时间变化模式,并且这些变量之间还存在着相互依赖的空间信息。为了充分提取这种时空特征,本文提出了基于融合时空特征的编码器-解码器(Spatio-Temporal EncoderDecoder,STED)航迹预测模型。在Encoder中使用门控循环单元(Gated Recurrent Unit,GRU)、卷积神经网络(Convolutional Neural Network,CNN)和注意力机制(Attention,AT)构成的双通道网络来分别提取航迹时空特征,Decoder对时空特征进行拼接融合,并利用GRU对融合特征进行学习和递归输出,实现对未来多步航迹信息的预测。利用真实的航迹数据对算法性能进行验证,实验结果表明,所提STED网络模型能够在未来10 min预测范围内进行高精度的短期航迹预测,相比于LSTM、CNN-LSTM和AT-LSTM等数据驱动航迹预测模型具有更高的精度。此外,STED网络模型预测一个航迹点平均耗时为0.002 s,具有良好的实时性。

Blind recognition of k/n rate convolutional encoders from noisy observation

Blind recognition of convolutional codes is not only essential for cognitive radio, but also for non-cooperative context. This paper is dedicated to the blind identification of rate k/n convolutional encoders in a noisy context based on Walsh-Hadamard transformation and block matrix (WHT-BM). The proposed algorithm constructs a system of noisy linear equations and utilizes all its coefficients to recover parity check matrix. It is able to make use of fault-tolerant feature of WHT, thus providing more accurate results and achieving better error performance in high raw bit error rate (BER) regions. Moreover, it is more computationally efficient with the use of the block matrix (BM) method. © 2017 Beijing Institute of Aerospace Information.

基于AutoEncoder的油气管道控制系统异常状态监测方法

压缩机控制电路的健康状态管理在管道运输中至关重要。通常油气管道压缩机系统部署地点远离城市,环境恶劣,且负荷高、工作时间长,因此故障频发。构建可靠的健康状态检测模型通常需要大量的故障样本,然而在实际数据中,故障样本相对稀缺。采用一种基于自编码器(auto encoder,AE)的单分类方法对油气管道控制系统的异常状态进行辨识。该模型仅需对系统的正常工作状态进行学习,通过编码器可实现特征的自适应提取,从而对数据进行抽象表示,并获得较好的非线性映射能力;当数据分布异常时,系统可区分其与正常信号间的差异,并进行预警。实验部分采用西部输油管道控制系统中实地获取的通信解码信号以及电源信号进行验证,并以单分类支持向量机方法作对比实验,表明了所提出方法的有效性。

基于Encoder-CNN的土壤氮含量光谱预测模型研究

基于光谱的土壤氮含量预测模型泛化能力弱是制约其推广应用的瓶颈。鉴于特征提取及非线性表达能力方面的优势,深度学习模型具有较强的泛化能力。提出一种融合自动编码器和卷积神经网络(Encoder-CNN)的土壤氮含量光谱预测模型,探索模型结构和参数对模型性能的影响。根据以往研究成果和相关性分析,获得180个与氮含量强相关的波长,将其作为Encoder-CNN模型输入,而将土壤氮含量作为模型输出。Encoder-CNN模型利用自动编码器的编码部分进行光谱数据降维,然后输入到卷积神经网络进行土壤氮含量预测。设计2种网络结构,每种网络结构包含2种不同参数设置,共4个模型,用以探索Encoder-CNN土壤氮含量光谱预测模型结构和参数对模型性能的影响。利用公开数据集LUCAS对模型进行训练。按3σ原则对公开数据集LUCAS进行异常值检测与处理,获得20791个数据,其中18711个样本作为训练集,2080个样本作为测试集,对Encoder-CNN模型进行训练。结果表明:对于自动编码器,在相同隐含层数下,最后的隐含层神经元个数为30时,复现效果最优。增加隐含层数,会提升复现效果。增加卷积核数量,特别是尺寸为1×1卷积核,能够提高模型的预测性能与可靠性。增加池化层的网络结构,模型预测精度提升至0.90以上。增加全连接层神经元数量也会提升模型性能。利用自采集的黑龙江黑土实时光谱数据集进行模型迁移,观察模型泛化能力。当模型迭代100次后,在黑龙江数据集上的预测精度即可达到0.90以上;当迭代次数为900时,模型在训练集和测试集上的预测精度可以达到0.98。结果表明,所构建的Encoder-CNN土壤氮含量光谱预测模型具有较好的泛化能力。

Damage Detection of Composite Material Intelligent Structure with a New Photoelectric System

A kind of photoelectric system that is suitable to measuring and to testing the damage of the composite material intelligent structure was presented. It can measure the degree of damage of the composite intelligent structure and it also can tell us the damage position in the structure. This system consists of two parts ： software and hardware. Experiments of the damage detection and the analysis of the composite material structure with the photoelectric system were performed, and a series of damage detection experiments was conducted. The results prove that the performance of the system is well and the effects of the measure and test are evident. Through all the experiments, the damage detection technology and test system are approved to be real-time, effective and reliable in the damage detection of the composite intelligent structure.

Upconversion Nanoparticles-Encoded Hydrogel Microbeads-Based Multiplexed Protein Detection

Fluorescently encoded microbeads are in demand for multiplexed applications in different fields.Compared to organic dye-based commercially available Luminex's x MAP technology, upconversion nanoparticles(UCNPs) are better alternatives due to their large antiStokes shift, photostability, nil background, and single wavelength excitation. Here, we developed a new multiplexed detection system using UCNPs for encoding poly(ethylene glycol) diacrylate(PEGDA) microbeads as well as for labeling reporter antibody. However, to prepare UCNPs-encoded microbeads, currently used swellingbased encapsulation leads to non-uniformity, which is undesirable for fluorescence-based multiplexing. Hence,we utilized droplet microfluidics to obtain encoded microbeads of uniform size, shape, and UCNPs distribution inside. Additionally, PEGDA microbeads lack functionality for probe antibodies conjugation on their surface.Methods to functionalize the surface of PEGDA microbeads(acrylic acid incorporation, polydopamine coating)reported thus far quench the fluorescence of UCNPs. Here,PEGDA microbeads surface was coated with silica followed by carboxyl modification without compromising the fluorescence intensity of UCNPs. In this study, droplet microfluidics-assisted UCNPs-encoded microbeads of uniform shape, size, and fluorescence were prepared.Multiple color codes were generated by mixing UCNPs emitting red and green colors at different ratios prior to encapsulation. UCNPs emitting blue color were used to label the reporter antibody. Probe antibodies were covalently immobilized on red UCNPs-encoded microbeads for specific capture of human serum albumin(HSA) as a model protein. The system was also demonstrated for multiplexed detection of both human C-reactive protein(hCRP) and HSA protein by immobilizing anti-h CRP antibodies on green UCNPs.