一种基于类不平衡学习的情感分析方法

针对网络评论中普遍存在的负面评论较少而影响力却较大的类不平衡问题, 提出一种基于类不平衡学习的情感分析方法. 该方法利用深度学习训练过程中的概率输出, 以计算样例的信息熵作为影响因子构建交叉信息熵损失函数. 在IMDB公开数据集上进行实验验证的结果表明, 基于集成信息熵损失函数的双向长短期记忆网络能处理类不平衡问题; 对数据的统计分析结果表明, 该策略能提升基于双向长短期记忆网络的评论情感极性分类性能. 针对AUC(area under curve)指标, 使用集成信息熵损失函数的双向长短期记忆网络模型比未考虑类不平衡的深度学习模型在中位数上最多提升15.3%.     

“二三结合”开发模式渗流特征研究

L油田B区块开展"二三结合"实验区,"二三结合"开发模式渗流特征对区块的合理高效开发具有重要意义。为了明确其渗流特征,利用油藏数值模拟软件Eclipse建立代表意义的油藏数值模型,研究区块"二三结合"开发模式下原射孔层和补孔层的渗流特征,分析了"二三结合"开发模式下,不同开发阶段综合含水率、采出程度、井底压力、地层压力及含水饱和度的变化特征。研究表明,"二三结合"开发模式先通过水驱开发挖潜原射孔层剩余油,提高厚油层顶部剩余油的采出程度,再通过聚合物驱进一步提高开发效果,通过水驱与聚驱的综合作用,最终得到较好的开发效果。     

Defining G protein-coupled receptor peptide ligand expressomes and signalomes in human and mouse islets

Introduction

Islets synthesise and secrete numerous peptides, some of which are known to be important regulators of islet function and glucose homeostasis. In this study, we quantified mRNAs encoding all peptide ligands of islet G protein-coupled receptors (GPCRs) in isolated human and mouse islets and carried out in vitro islet hormone secretion studies to provide functional confirmation for the species-specific role of peptide YY (PYY) in mouse islets.

Materials and methods

GPCR peptide ligand mRNAs in human and mouse islets were quantified by quantitative real-time PCR relative to the reference genes ACTB, GAPDH, PPIA, TBP and TFRC. The pathways connecting GPCR peptide ligands with their receptors were identified by manual searches in the PubMed, IUPHAR and Ingenuity databases. Distribution of PYY protein in mouse and human islets was determined by immunohistochemistry. Insulin, glucagon and somatostatin secretion from islets was measured by radioimmunoassay.

Results

We have quantified GPCR peptide ligand mRNA expression in human and mouse islets and created specific signalomes mapping the pathways by which islet peptide ligands regulate human and mouse GPCR signalling. We also identified species-specific islet expression of several GPCR ligands. In particular, PYY mRNA levels were ~ 40,000-fold higher in mouse than human islets, suggesting a more important role of locally secreted Pyy in mouse islets. This was confirmed by IHC and functional experiments measuring insulin, glucagon and somatostatin secretion.

Discussion

The detailed human and mouse islet GPCR peptide ligand atlases will allow accurate translation of mouse islet functional studies for the identification of GPCR/peptide signalling pathways relevant for human physiology, which may lead to novel treatment modalities of diabetes and metabolic disease.

     

The modernity of Dedekind’s anticipations contained in What are numbers and what are they good for?

Archive for History of Exact Sciences - We show that Dedekind, in his proof of the principle of definition by mathematical recursion, used implicitly both the concept of an inductive cone from an...     

普通话在线游戏学习系统的研究与应用

针对希望进一步掌握普通话口语学习的外国学习者,该研究提出了一种普通话游戏学习系统.系统的学习模式结合了计算机技术和语言学研究成果,打破了传统的教学模式的限制,对学习者进行实时评测,提供学习报告与口语学习的最终评定结果.结果表明,该系统能够较为有效地辅助外国学习者的普通话口语学习.     

The expansion of GPCR transactivation-dependent signalling to include serine/threonine kinase receptors represents a new cell signalling frontier

G protein-coupled receptor (GPCR) signalling is mediated through transactivation-independent signalling pathways or the transactivation of protein tyrosine kinase receptors and the recently reported activation of the serine/threonine kinase receptors, most notably the transforming growth factor-β receptor family. Since the original observation of GPCR transactivation of protein tyrosine kinase receptors, there has been considerable work on the mechanism of transactivation and several pathways are prominent. These pathways include the “triple membrane bypass” pathway and the generation of reactive oxygen species. The recent recognition of GPCR transactivation of serine/threonine kinase receptors enormously broadens the GPCR signalling paradigm. It may be predicted that the transactivation of serine/threonine kinase receptors would have mechanistic similarities with transactivation of tyrosine kinase pathways; however, initial studies suggest that these two transactivation pathways are mechanistically distinct. Important questions are the relative importance of tyrosine and serine/threonine transactivation pathways, the contribution of transactivation to overall GPCR signalling, mechanisms of transactivation and the range of cell types in which this phenomenon occurs. The ultimate significance of transactivation-dependent signalling remains to be defined but it appears to be prominent and if so will represent a new cell signalling frontier.