Impact of viscous dissipation and Dufour on MHD natural convective flow past an accelerated vertical plate with Hall current

The present research work concentrates on viscous dissipation, Dufour, and heat source on an unsteady magnetohydrodynamics natural convective flow of a viscous, incompressible, and electrically conducting fluid past an exponentially accelerated infinite vertical plate in the existence of a strong magnetic field. The presence of the Hall current induces a secondary flow in the problem. The distinguishing features of viscous dissipation and heat flux produced due to gradient of concentration included in the model along with heat source as they are known to arise in thermal-magnetic polymeric processing. The flow equations are discretized implicitly using the finite difference method and solved using MATLAB fsolve routine. Numerical values of the primary and secondary velocities, temperature, concentration, skin friction, Nusselt number, and Sherwood number are illustrated and presented via graphs and tables for various pertinent parametric values. The Dufour effect was observed to strengthen the velocity and temperature profile in the flow domain. In contrast, due to the impact of viscous dissipation, the local Nusselt number reduces. The study also reveals that the inclusion of the chemical reaction term augments the mass transfer rate and diminishes the heat transfer rate at the plate.     

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.     

Preparation of faceted TaN grains by heating Ta-containing oxides in BN crucible with Na

Single-crystal grains of TaN were synthesized by heating Ta₂O₅, FeTa₂O₆, or FeTaO₄ in a BN crucible together with Na metal in an Ar atmosphere at 1100 °C. The BN crucible acted a solid source of nitrogen. Aggregates of columnar ε-TaN single crystals 10–150 μm in size were formed on the inner wall of the BN crucible when either Ta₂O₅ or FeTa₂O₆ was used. On the other hand, platelet-like single crystals of θ-TaN 1–50 μm in size were obtained from FeTaO₄. The results of wavelength-dispersive X-ray spectrometry indicated that the compositions of the ε-TaN and θ-TaN crystals were close to the stoichiometric ideal.     

结合预训练模型和语言知识库的文本匹配方法

针对文本匹配任务,该文提出一种大规模预训练模型融合外部语言知识库的方法。该方法在大规模预训练模型的基础上,通过生成基于WordNet的同义—反义词汇知识学习任务和词组—搭配知识学习任务引入外部语言学知识。进而,与MT-DNN多任务学习模型进行联合训练,以进一步提高模型性能。最后利用文本匹配标注数据进行微调。在MRPC和QQP两个公开数据集的实验结果显示,该方法可以在大规模预训练模型和微调的框架基础上,通过引入外部语言知识进行联合训练有效提升文本匹配性能。     

面向智能客服系统的情感分析技术

该文以阿里小蜜为例,对智能客服系统中的情感分析技术进行比较全面的介绍,包括情感分析算法模型的原理及其在智能客服系统的多个应用场景中的实际落地使用方式和效果分析。智能客服在解决客户高频业务问题的同时,也需要给客户提供多维度的、具有类人能力的助理、导购、语聊和娱乐等服务,提高客户对智能客服机器人的整体满意度。在此过程中,情感分析技术在机器人类人能力建设中起到了至关重要的作用。该文围绕智能客服系统中人机结合的服务形式,从六个维度总结和介绍了情感分析技术在智能客服系统中的应用场景,包括用户情感检测、用户情感安抚、情感生成式语聊、客服服务质检、会话满意度预估和智能人工入口。     

光电直读光谱仪在管线钢成分分析中的应用

为了进一步提高SPECTRO LAB直读光谱仪在管线钢成分分析中的准确度和效率，通过参数调整及控样研究，对直读光谱仪分析方法进行了优化，并提出了两种优化方案。结果表明，调整光源参数，能够缩短试验分析时间，减少氩气消耗；综合控样的研究实施，可以解决多块标样分步骤操作的繁琐和误差，省时省力。经实践验证，优化后的方案能够满足生产检验要求，分析方法简单、快捷，降低了成本消耗。     

基于JJG(闽)1086-2018放射治疗CT模拟定位机X射线辐射源的质控分析

整个精确放疗流程最关键、最基础的一个环节就是定位CT扫描精确程度,做好定位CT的质控关乎整个放疗流程的成败。介绍了JJG(闽)1086-2018放射治疗CT模拟定位机X射线辐射源在检定和质控中的流程和方法,从剂量性能要求、图像成像质量和机器机械精度方面对CT模拟定位机的质量控制做了定量的要求。连续8个月的测试结果表明:CT测量值误差在30HU以内,图像均匀性≤5.0HU,空间分辨力7Lp/cm,低对比度分辨力在对比度为1.0%,0.5%,0.3%时分别≤3.0mm,≤5.0mm,≤6.0mm,达到了规程的要求。CT模拟定位机的图像质量和机械精确程度得到了很好的保障,定位CT的剂量准确,影像参数和位置精确度有规可循。     

Design and optimization of radioisotope sources for betavoltaic batteries

A Monte Carlo source model using PENELOPE was developed to investigate different tritiated metals in order to design a better radioisotope source for betavoltaic batteries. The source model takes into account the self‐absorption of beta particles in the source which is a major factor for an efficient source design. The average beta energy, beta flux, source power output, and source efficiency were estimated for various source thicknesses. The simulated results for titanium tritide with 0° and 90° angular distributions of beta particles were validated with experimental results. The importance of the backscattering effect due to isotropic particle emission was analyzed. The results showed that the normalized average beta energy increases with the source thickness, and it reaches peak energy depending on the density and the specific activity of the source. The beta flux and power output also increase with increasing source thickness. However, the incremental increase in beta flux and power output becomes minimal for higher thicknesses, as the source efficiency decreases significantly at higher thicknesses due to the self‐absorption effect. Thus, a saturation threshold is reached. A low‐density source material such as beryllium tritide provided a higher power output with higher efficiency. A maximum power output of approximately 4 mW/cm³ was obtained for beryllium tritide with SiC. A form factor approach was used to estimate the optimum source thickness. The optimum source thickness was found near the thickness where the peak beta particle average energy occurs.     

Biomarker geochemistry of Eq34-10 cyclothem shale in Qianjiang Depression of the Jianghan salt lake facies Basin

The objective of this study is to provide biomarker geochemistry information on shale between sheet-like salts from Qianjiang Depression of the Jianghan salt lake facies Basin, and illustrate the depositional conditions, source organic matter input and maturation. 7 shale samples from Eq3⁴-10 rhythmic were analysed with gas chromatography-mass spectrometry. The investigated biomarkers show that the shale oil are charactertized by an apperent odd-even preference of n-alkane, low Pr/Ph values and a high content of gammacerane and C₃₀+ hopanes, maderate C_21-/C₂₁₊ ratios, the high relative content of C₂₇ 5α(H), 14α(H), 17α(H)- steranes (20R), and an anomaly high C₂₉ sterane ββ/(ββ + αα) value. These biomarker geochemistry characterizations suggested that the shale oils were derived from mixed aquatic organisms and land plants, and deposited in an anoxic and saline-hypersaline depositional environment.     

基于多元神经网络融合的分布式资源空间文本分类研究

针对实体产业对科技资源的服务需求,以服务效应作为资源文本分类标准,提出一种基于多元神经网络融合的分布式资源空间文本分类模型。设计了包含词嵌入层、卷积层、双向门控循环单元层、注意力机制层和softmax层的多元神经网络通路;在此基础上采用基于需求—效应—资源分类策略,完成了从定性科技资源需求到定量资源服务效应求解,再到定性科技资源输出的映射变换,重点解决了分布式科技资源局部和全局语义特征形式多样、文本长距离依赖特征显著、重要资源信息难以准确识别的问题,进而从分布式科技资源空间中快速准确地获取效应知识,提升实体产业产品研发效率和创新能力;通过万方专利科技资源数据集验证了所提方法的可行性和有效性,为更加全面地挖掘资源文本特征和按需服务实体产业提供了一种新的思路和手段。