Few-shot node classification via local adaptive discriminant structure learning

Node classification has a wide range of application scenarios such as citation analysis and social network analysis. In many real-world attributed networks, a large portion of classes only contain limited labeled nodes. Most of the existing node classification methods cannot be used for few-shot node classification. To train the model effectively and improve the robustness and reliability of the model with scarce labeled samples, in this paper, we propose a local adaptive discriminant structure learning (LADSL) method for few-shot node classification. LADSL aims to properly represent the nodes in the attributed graphs and learn a metric space with a strong discriminating power by reducing the intra-class variations and enlarging inter-class differences. Extensive experiments conducted on various attributed networks datasets demonstrate that LADSL is superior to the other methods on few-shot node classification task.     

Complex Attributed Network Embedding for medical complication prediction

Knowledge and Information Systems - To assure the development of effective treatment plans, it is crucial for understanding the complication relationships among diseases. In practice, traditional...     

Decomposed-distance weighted optimal transport for unsupervised domain adaptation

Applied Intelligence - Unsupervised Domain Adaptation (UDA) aims to transfer knowledge from a label-rich source domain to an unlabeled target domain with a different but related distribution....     

Characteristic analysis of Fe3O4 nanoparticles modified by oleic acid and undecylenic acid

Carbonization of magnetic polymer microspheres is one of the methods for the preparation of magnetic carbon materials. Fe₃O₄ magnetic particle characteristics considerably influence the magnetic content and size distribution of magnetic polymer microspheres. The characteristics of Fe₃O₄ nanoparticles modified by oleic acid (OA) and undecylenic acid (UA) were analyzed by X-ray diffraction, Fourier transform infrared, scanning electron microscopy, dynamic laser light scattering, thermogravimetry/differential thermogravimetry, vibrating sample magnetometer, and water contact angle. Fe₃O₄ nanoparticles modified by OA and UA are nearly spherical and exhibit superparamagnetism. Fe₃O₄ particle size and saturation magnetization are slightly influenced by the OA and UA composition. OA and UA both are chemically adsorbed onto Fe₃O₄ as bidentate chelates. OA shows easier adsorption onto Fe₃O₄ than UA. OA groups have an expanded arrangement on OA@Fe₃O₄, whereas UA groups have a condensed arrangement on UA@Fe₃O₄. Particle lipophilicity decreases and particle clustering increases with decreasing OA content and increasing UA content on OA-UA@Fe₃O₄ nanoparticles.     

国产800级碳纤维表面状态及其复合材料界面性能EI北大核心CSCD

通过扫描电子显微镜（SEM）、原子力显微镜（AFM）、X射线光电子能谱（XPS）、复丝拉伸法、单丝拉伸法及单丝断裂法对3种国产800-12K碳纤维表面状态及其复丝拉伸性能、单丝复合体系的界面性能进行系统分析与研究。结果表明：3种国产800级碳纤维表面均较为光滑,纤维的粗糙度为9-17nm,纤维表面含氧量较高且稳定,O/C在0.23~0.27之间;3种国产800级碳纤维复丝拉伸强度相当,质量控制稳定,断裂伸长率为1.9左右,纤维与树脂基体匹配性较好;3种国产800级碳纤维单丝拉伸强度不稳定,纤维的表面化学活性对纤维与树脂基体的界面结合强度影响显著。     

DOC‐LS,a new liposome for dermal delivery,and its endocytosis by HaCaT and CCC‐ESF‐1 cells

The main objective of this work was to investigate the uptake channels of skin cells through which coumarin 6, transported by deoxycholate‐mediated liposomes (DOC‐LS), was internalised; this was also compared against the action of conventional LS. Coumarin 6‐loaded DOC‐LS and LS were characterised for size distribution, zeta potential, and shape, and analysed in vitro in human epidermal immortal keratinocyte (HaCaT) (epidermal) and human embryonic skin fibroblast (CCC‐ESF‐1) (dermal) cell lines. Various endocytosis inhibitors were incubated with cells treated with the nanocarriers. Flow cytometry results indicated that HaCaT and CCC‐ESF‐1 cells internalise the tested preparations through pinocytotic vesicles, macropinocytosis, clathrin‐mediated endocytic pathways, and via lysosomes, which consume a considerable amount of energy. The endocytosis pathways of DOC‐LS and LS showed no difference. This study provides a basis for the application of LS being combined with a microneedle system for efficient intracellular drug delivery, targeting cutaneous histocyte disorders.Inspec keywords: drugs, nanoparticles, lipid bilayers, nanomedicine, biomedical materials, electrokinetic effects, biomembrane transport, drug delivery systems, skin, organic compoundsOther keywords: dermal delivery, CCC‐ESF‐1 cells, skin cells, deoxycholate‐mediated liposomes, coumarin 6‐loaded DOC‐LS, endocytosis inhibitors, clathrin‐mediated endocytic pathways, endocytosis pathways, HaCaT cell lines, size distribution, zeta potential, nanocarriers, flow cytometry, pinocytotic vesicles, macropinocytosis, microneedle system, efficient intracellular drug delivery, targeting cutaneous histocyte disorders     

Controllable defects implantation in MoS<Subscript>2</Subscript> grown by chemical vapor deposition for photoluminescence enhancement

Photoluminescence (PL) of transition metal dichalcogenides (TMDs) can be engineered by controlling the density of defects, which provide active sites for electron-hole recombination, either radiatively or non-radiatively. However, the implantation of defects by external stimulation, such as uniaxial tension and irradiation, tends to introduce local damages or structural non-homogeneity, which greatly degrades their luminescence properties and impede their applicability in constructing optoelectronic devices. In this paper, we present a strategy to introduce a controllable level of defects into the MoS₂ monolayers by adding a hydrogen flow during the chemical vapor deposition, without sacrificing their luminescence characteristics. The density of the defect is controlled directly by the concentration of hydrogen. For an appropriate hydrogen flux, the monolayer MoS₂ sheets have three times stronger PL emission at the excitonic transitions, compared with those samples with nearly perfect crystalline structure. The defect-bounded exciton transitions at lower energies arising in the defective samples and are maximized when the total PL is the strongest. However, the B exciton, exhibits a monotonic decline as the defect density increases. The Raman spectra of the defective MoS₂ reveal a redshift (blueshift) of the in-plane (out-of-plane) vibration modes as the hydrogen flux increases. All the evidence indicates that the generated defects are in the form of sulfur vacancies. This study renders the high-throughput synthesis of defective MoS₂ possible for catalysis or light emitting applications.

Open image in new window

     

基于数字兆欧表自动切换档位的研究与实现

随着电力、电气设备的日益复杂化,对于其安全性能也有了更高的要求,数字兆欧表作为测量绝缘阻值的重要工具,已在工业领域中广泛应用;为了满足测量精度以及兆欧表的可靠性,通过基于S3C2440的嵌入式平台,提出了一种基于数字兆欧表的自动切换档位的实现方案,以达到更加精确的测量;此方案通过判断采集到的电压值选择相应的采样电阻来实现,并结合硬件平台和软件设计,分别介绍了驱动程序和应用程序的编写与测试,并很好地应用到了实际电路中;经测试,能够达到预期的效果,测量精度高,测量误差控制在±2%。     

D-S算法在粮情监控信息融合系统中的应用

本文通过多传感器对影响储藏粮食因子的检测,利用D-S融合算法,得到较为科学、合理的储藏粮食影响因素的判断,为在储藏粮食过程中采取适当的措施提供了强有力的依据,为进一步评估储粮品质打下基础。     

基于无线传感器网络的管道检测精确定位

为了对大范围的工业管道检测装置进行精确定位,采用无线传感器网络,在工业管道外布置磁传感器模块来测量管道检测装置的磁信号.每个磁传感器模块由磁传感单元、处理单元、收发单元、电源单元等功能模块组成.磁传感器模块的测量数据由无线传感器网络传送到处理中心.处理中心依据无线传感器网络发回的磁信号数据,由磁定位算法计算得到管道检测装置所处的精确位置,从而实现管道检测位置的精确定位.