网络链接预测能够获取网络中丢失链接的重要信息或进行网络的动态演变分析.现有的基于节点相似性的网络链接预测方法往往针对简单的一(多)阶邻居信息或特定类型的小型网络,设计较为复杂的计算方法,其扩展性和大规模网络中的可计算性都...网络链接预测能够获取网络中丢失链接的重要信息或进行网络的动态演变分析.现有的基于节点相似性的网络链接预测方法往往针对简单的一(多)阶邻居信息或特定类型的小型网络,设计较为复杂的计算方法,其扩展性和大规模网络中的可计算性都受到了严峻的挑战.文中基于深度学习在神经网络语言模型中应用的启发,提出了一个LsNet2Vec(Large-scale Network to Vector)模型.通过结合随机游走的网络数据集序列化方法,进行大规模的无监督机器学习,从而将网络中节点的结构特征信息映射到一个连续的、固定维度的实数向量.然后,使用学习到的节点结构特征向量,就可以迅速计算大规模网络中任意节点之间的相似度,以此来进行网络中的链接预测.通过在16个大规模真实数据集上和目前的多个基准的最优预测算法对比发现,LsNet2Vec模型所得到的预测总体效果是最优的:在保证了大规模网络中链接预测计算可行性的同时,于多个数据集上相对已有方法呈现出较大的AUC值提升,最高达8.9%.展开更多
In order to explain the phenomenon of negative capacitance（NC） in light emitting diodes LEDs, we present a new model based on local strong recombination in active region and firstly deduce the analytic expression of...In order to explain the phenomenon of negative capacitance（NC） in light emitting diodes LEDs, we present a new model based on local strong recombination in active region and firstly deduce the analytic expression of NO from continuity equation. The theoretical result indicates that the NC effect becomes stronger when the carrier recombination rate increases in a certain range,which is consistent with the experimental result. Accordingly,we confirm that the NO is caused by carrier recombination in active reaion instead of by other exterior factors.展开更多
Objective: To monitor the systemic gene expression profile in a murine model of lipopolysaccharide-induced acute lung injury. Methods: Acute lung injury was induced by intratracheal injection of lipopolysaccharide in ...Objective: To monitor the systemic gene expression profile in a murine model of lipopolysaccharide-induced acute lung injury. Methods: Acute lung injury was induced by intratracheal injection of lipopolysaccharide in 3 mice. Another 3 normal mice receiving same volume of normal saline were taken as the controls. The comprehensive gene expression profile was monitored by the recently modified long serial analysis of gene expression. Results: A total of 24 670 tags representing 12 168 transcripts in the control mice and 26 378 tags representing 13 397 transcripts in the mice with lung injury were identified respectively. There were 11 transcripts increasing and 7 transcripts decreasing more than 10 folds in the lipopolysaccharide-treated mice. The most overexpressed genes in the mice with lung injury included serum amyloid A3, metallothionein 2, lipocalin 2, cyclin-dependent kinase inhibitor 1A, lactate dehydrogenase 1, melatonin receptor, S100 calcium-binding protein A9, natriuretic peptide precursor, etc. Mitogen activated protein kinase 3, serum albumin, complement component 1 inhibitor, and ATP synthase were underexpressed in the lung injury mice. Conclusions: Serial analysis of gene expression provides a molecular characteristic of acute lung injury.展开更多
文摘网络链接预测能够获取网络中丢失链接的重要信息或进行网络的动态演变分析.现有的基于节点相似性的网络链接预测方法往往针对简单的一(多)阶邻居信息或特定类型的小型网络,设计较为复杂的计算方法,其扩展性和大规模网络中的可计算性都受到了严峻的挑战.文中基于深度学习在神经网络语言模型中应用的启发,提出了一个LsNet2Vec(Large-scale Network to Vector)模型.通过结合随机游走的网络数据集序列化方法,进行大规模的无监督机器学习,从而将网络中节点的结构特征信息映射到一个连续的、固定维度的实数向量.然后,使用学习到的节点结构特征向量,就可以迅速计算大规模网络中任意节点之间的相似度,以此来进行网络中的链接预测.通过在16个大规模真实数据集上和目前的多个基准的最优预测算法对比发现,LsNet2Vec模型所得到的预测总体效果是最优的:在保证了大规模网络中链接预测计算可行性的同时,于多个数据集上相对已有方法呈现出较大的AUC值提升,最高达8.9%.
基金This work was supported by the National Nature Science Foun-dation (Grant No.DMR-60376027) .
文摘In order to explain the phenomenon of negative capacitance（NC） in light emitting diodes LEDs, we present a new model based on local strong recombination in active region and firstly deduce the analytic expression of NO from continuity equation. The theoretical result indicates that the NC effect becomes stronger when the carrier recombination rate increases in a certain range,which is consistent with the experimental result. Accordingly,we confirm that the NO is caused by carrier recombination in active reaion instead of by other exterior factors.
文摘Objective: To monitor the systemic gene expression profile in a murine model of lipopolysaccharide-induced acute lung injury. Methods: Acute lung injury was induced by intratracheal injection of lipopolysaccharide in 3 mice. Another 3 normal mice receiving same volume of normal saline were taken as the controls. The comprehensive gene expression profile was monitored by the recently modified long serial analysis of gene expression. Results: A total of 24 670 tags representing 12 168 transcripts in the control mice and 26 378 tags representing 13 397 transcripts in the mice with lung injury were identified respectively. There were 11 transcripts increasing and 7 transcripts decreasing more than 10 folds in the lipopolysaccharide-treated mice. The most overexpressed genes in the mice with lung injury included serum amyloid A3, metallothionein 2, lipocalin 2, cyclin-dependent kinase inhibitor 1A, lactate dehydrogenase 1, melatonin receptor, S100 calcium-binding protein A9, natriuretic peptide precursor, etc. Mitogen activated protein kinase 3, serum albumin, complement component 1 inhibitor, and ATP synthase were underexpressed in the lung injury mice. Conclusions: Serial analysis of gene expression provides a molecular characteristic of acute lung injury.
