融合MAML和CBAM的安卓恶意应用家族分类模型

为满足对新兴安卓恶意应用家族的快速检测需求,提出一种融合MAML(model-agnostic meta-learning)和CBAM(convolutional block attention module)的安卓恶意应用家族分类模型MAML-CAS。将安卓恶意应用样本集中的DEX文件可视化为灰度图,并构建任务集;融合混合域注意力机制CBAM,设计两个具有同等结构的卷积神经网络,分别作为基学习器和元学习器,这两个学习器在自动提取任务集中样本特征的同时,可从通道和空间两个维度来增强关键特征表达;利用元学习方法 MAML对两个学习器进行训练,其中基学习器完成特定恶意家族分类任务的属性学习,元学习器则学习不同任务的共性;在两个学习器训练完成后,MAML-CAS将获得初始化参数,在面对新的安卓恶意应用家族分类任务时,不需要重新训练,只需要少量样本就可以快速迭代;利用训练完成的基学习器提取安卓恶意应用家族特征,并利用SVM进行恶意家族分类。实验结果表明,MAML-CAS模型对新兴小样本安卓恶意应用家族具有良好的检测效果,检测速度较快,并具有较好的稳定性。     

Fast medical concept normalization for biomedical literature based on stack and index optimized self-attention

Neural Computing and Applications - Medical concept normalization aims to construct a semantic mapping between mentions and concepts and to uniformly represent mentions that belong to the same...     

Sensitive Detection of Carcinoembryonic Antigen Using Stability‐Limited Few‐Layer Black Phosphorus as an Electron Donor and a Reservoir

The instability of few‐layer black phosphorus (FL‐BP) hampers its further applications. Here, it can be demonstrated that the instability of FL‐BP can also be the advantage for application in biosensor. First, gold nanoparticle/FL‐BP (BP‐Au) hybrid is facilely synthesized by mixing Au precursor with FL‐BP. BP‐Au shows outstanding catalytic activity (K = 1120 s^?1 g^?1) and low activation energy (17.53 kJ mol^?1) for reducing 4‐nitrophenol, which is attributed to the electron‐reservoir and electron‐donor properties of FL‐BP, and synergistic interaction of Au nanoparticles and FL‐BP. Oxidation of FL‐BP after catalytic reaction is further confirmed by transmission electron microscope, X‐ray photoelectron spectroscopy, and zeta potentials. Second, the catalytic activity of BP‐Au can be reversibly switched from “inactive” to “active” upon treatment with antibody and antigen in solution, thus providing a versatile platform for label‐free colorimetric detection of biomarkers. The sensor shows a wide detection range (1 pg mL^?1 to –10 µg mL^?1), high sensitivity (0.20 pg mL^?1), and selectivity for detecting carcinoembryonic antigen (CEA). Finally, the biosensor has been used to detect CEA in colon and breast cancer clinical samples with satisfactory results. Therefore, the instability of BP can also be the advantage for application in detecting cancer biomarker in clinic.     

气相法白炭黑填充天然橡胶的裂纹扩展机理

研究了不同粒径,不同份数的白炭黑对天然橡胶(NR)裂纹生长行为的影响。采用具有裂纹生长测试功能的DMA(DMA-Crack Growth)对预割口试样的裂纹扩展速度进行测试。利用超高速显微摄像技术对裂纹扩展尖端进行在线追踪,得到裂纹扩展尖端的动态显微图像。结果表明,添加20 phr原生粒径为15 nm的气相法白炭黑时,裂纹尖端韧带均匀且细密,对裂纹生长起到很好的阻碍作用;添加20 phr原生粒径为7 nm的气相法白炭黑时,对试样的裂纹生长有不利影响。     

电气检测领域电源稳定性核查方法

产品检测中电源具有十分重要的作用,本文以电动机检测实验室为例,介绍了电器领域中电源稳定性的核查方法。