A Smart Deep Convolutional Neural Network for Real-Time Surface Inspection

A proper detection and classification of defects in steel sheets in real time have become a requirement for manufacturing these products, largely used in many industrial sectors. However, computers used in the production line of small to medium size companies, in general, lack performance to attend real-time inspection with high processing demands. In this paper, a smart deep convolutional neural network for using in real-time surface inspection of steel rolling sheets is proposed. The architecture is based on the state-of-the-art SqueezeNet approach, which was originally developed for usage with autonomous vehicles. The main features of the proposed model are: small size and low computational burden. The model is 10 to 20 times smaller when compared to other networks designed for the same task, and more than 700 times smaller than general networks. Also, the number of floating-point operations for a prediction is about 50 times lower than the ones used for similar tasks. Despite its small size, the proposed model achieved near-perfect accuracy on a public dataset of 1800 images of six types of steel rolling defects.     

Leucokinin and Associated Neuropeptides Regulate Multiple Aspects of Physiology and Behavior in Drosophila

Leucokinins (LKs) constitute a family of neuropeptides identified in numerous insects and many other invertebrates. LKs act on G-protein-coupled receptors that display only distant relations to other known receptors. In adult Drosophila, 26 neurons/neurosecretory cells of three main types express LK. The four brain interneurons are of two types, and these are implicated in several important functions in the fly’s behavior and physiology, including feeding, sleep–metabolism interactions, state-dependent memory formation, as well as modulation of gustatory sensitivity and nociception. The 22 neurosecretory cells (abdominal LK neurons, ABLKs) of the abdominal neuromeres co-express LK and a diuretic hormone (DH44), and together, these regulate water and ion homeostasis and associated stress as well as food intake. In Drosophila larvae, LK neurons modulate locomotion, escape responses and aspects of ecdysis behavior. A set of lateral neurosecretory cells, ALKs (anterior LK neurons), in the brain express LK in larvae, but inconsistently so in adults. These ALKs co-express three other neuropeptides and regulate water and ion homeostasis, feeding, and drinking, but the specific role of LK is not yet known. This review summarizes Drosophila data on embryonic lineages of LK neurons, functional roles of individual LK neuron types, interactions with other peptidergic systems, and orchestrating functions of LK.     

Sequential alignment attention model for scene text recognition

Scene text recognition has been a hot research topic in computer vision due to its various applications. The state-of-the-art solutions usually depend on the attention-based encoder-decoder framework that learns the mapping between input images and output sequences in a purely data-driven way. Unfortunately, there often exists severe misalignment between feature areas and text labels in real-world scenarios. To address this problem, this paper proposes a sequential alignment attention model to enhance the alignment between input images and output character sequences. In this model, an attention gated recurrent unit (AGRU) is first devised to distinguish the text and background regions, and further extract the localized features focusing on sequential text regions. Furthermore, CTC guided decoding strategy is integrated into the popular attention-based decoder, which not only helps to boost the convergence of the training but also enhances the well-aligned sequence recognition. Extensive experiments on various benchmarks, including the IIIT5k, SVT, and ICDAR datasets, show that our method substantially outperforms the state-of-the-art methods.     

Hepatitis B Virus-Like Particle: Targeted Delivery of Plasmid Expressing Short Hairpin RNA for Silencing the Bcl-2 Gene in Cervical Cancer Cells

Gene therapy research has advanced to clinical trials, but it is hampered by unstable nucleic acids packaged inside carriers and there is a lack of specificity towards targeted sites in the body. This study aims to address gene therapy limitations by encapsidating a plasmid synthesizing a short hairpin RNA (shRNA) that targets the anti-apoptotic Bcl-2 gene using truncated hepatitis B core antigen (tHBcAg) virus-like particle (VLP). A shRNA sequence targeting anti-apoptotic Bcl-2 was synthesized and cloned into the pSilencer 2.0-U6 vector. The recombinant plasmid, namely PshRNA, was encapsidated inside tHBcAg VLP and conjugated with folic acid (FA) to produce FA-tHBcAg-PshRNA VLP. Electron microscopy revealed that the FA-tHBcAg-PshRNA VLP has an icosahedral structure that is similar to the unmodified tHBcAg VLP. Delivery of FA-tHBcAg-PshRNA VLP into HeLa cells overexpressing the folate receptor significantly downregulated the expression of anti-apoptotic Bcl-2 at 48 and 72 h post-transfection. The 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay demonstrated that the cells’ viability was significantly reduced from 89.46% at 24 h to 64.52% and 60.63%, respectively, at 48 and 72 h post-transfection. As a conclusion, tHBcAg VLP can be used as a carrier for a receptor-mediated targeted delivery of a therapeutic plasmid encoding shRNA for gene silencing in cancer cells.     

投影寻踪分类模型在常见造纸纤维原料综合评价中的应用

将投影寻踪分类（PPC）模型与基于实数编码的加速遗传算法（RAGA）相结合，同时优化多个指标参数，将高维数据指标转化为低维空间上的一维投影值，建立RAGA-PPC模型，用于造纸纤维原料分类，并对造纸纤维原料进行综合评价。结果表明，基于RAGA-PPC模型的评价结果与造纸纤维原料实际分类结果一致，此方法客观可靠，精度高，具有一定的应用前景。     

计及RSC控制的DFIG短路电流直流分量解析表达

现有文献针对计及转子变流器(RSC)控制的双馈感应风电机组(DFIG)定子短路电流解析表达,将定子磁链当作一阶直流衰减分量或忽略功率外环控制。基于DFIG电压、磁链和RSC控制方程,得到定子电流关于定子电压和定子功率的传递函数,提出定子电流的精确解析表达式。基于RSC内、外环PI参数关系,推导直流分量衰减时间常数和角频率关于PI参数的表达式。分析了RSC内外环PI参数对定子电流直流衰减分量的影响。仿真结果验证了解析表达式的准确性,为PI参数选取和保护装置测量、整定提供依据。     

基于全相位预处理的时域多重信号分类波达方向估计方法

低信噪比下,针对宽带短脉冲情况下频域多重信号分类(MUSIC)中噪声子空间估计不稳定问题,提出一种基于全相位预处理的时域多重信号分类波达方向(DOA)估计方法。①对线列阵接收数据进行分组处理;②按搜索角度对各组数据进行相移预处理,并对各组数据预处理结果进行相加,得到一组新数据;③对线列阵接收数据在时域构建相移后的协方差矩阵,在更短数据长度下,稳定实现噪声子空间估计,并依据估计出的噪声子空间含有的正交特性,通过单位矩阵加法器得到相应空间谱估计值,实现波达方向估计。数值仿真和实测数据处理结果表明,相比频域MUSIC方法,该方法有效提高了线列阵接收数据协方差矩阵中信号含有量和信噪比,能够在更短数据长度情况下实现对噪声子空间的稳定估计,具有较好的稳定性和检测性能,提高了MUSIC方法在实际波达方向估计中的鲁棒性。     

带权决策表的属性约简

属性约简是粗糙集理论的重要应用。考虑将决策表中的每行都作为一条决策规则时，若把表中出现相同决策规则的次数作为权，可得到带权决策表。提出了关于带权决策表的正域约简相应的辨识矩阵并给出了证明，从而得到了约简算法。相比于决策表中的正域约简时发现，通过将决策表转化为带权决策表后，再利用算法1进行约简时，其在一定程度上优于前者。提出了近似分类精度约简相应的辨识矩阵并给出了证明。对于2个算法，在选取的UCI数据集上进行了实验验证。通过实验进一步说明了所提出算法的可行性和有效性。     

基于改进的Random Subspace的客户投诉分类方法

电信业的客户投诉不断增多而又亟待高效处理。针对电信客户投诉数据的特点，提出了一种面向高维数据的改进的集成学习分类方法。该方法综合考虑客户投诉中的文本信息及客户通讯状态信息，基于Random Subspace方法，以支持向量机（Support Vector Machine，SVM）为基分类器，采用证据推理（Evidential Reasoning，ER）规则为一种新的集成策略，构造分类模型对电信客户投诉进行分类。所提模型和方法在某电信公司客户投诉数据上进行了验证，实验结果显示该方法能够显著提高客户投诉分类的准确率和投诉处理效率。