基于细胞图卷积的组织病理图像分类研究

针对传统CNN（Convolutional Neural Network）在组织病理图像分类中存在的两个问题：其一,受限于内存大小,CNN无法对高分辨率的病理图像直接进行训练,这不可避免地破坏了细胞之间的空间结构信息,且无法学习全局的特征信息;其二,病理图像中的正常细胞和癌变细胞均有自身的病理学图像特征并且在空间上具有一定的关联性,但在结构化的二维阵列图像中无法被充分的表达。提出一种基于细胞图卷积（Cell-Graph Convolutional Network,C-GCN）的组织病理图像分类方法,将高分辨率的病理图像转换为图结构,在传统的GCN中将GraphSAGE（Graph SAmple and aggregate）模块与图池化相结合,提取出更具有代表性的一般性特征,使得C-GCN可以直接在高分辨率的组织学图像中学习特征,提高了模型的鲁棒性。     

Self-attention Hypergraph Pooling Network

Recently, Graph Convolutional neural Networks (GCNs) have attracted much attention by generalizing convolutional neural networks to graph data, which includes redefining convolution and pooling operations on graphs. Due to the limitation that graph data can only focus on dyadic relations, it cannot perform well in real practice. In contrast, a hypergraph can capture high-order data interaction and is easy to deal with complex data representation using its flexible hyperedges. However, the existing methods for hypergraph convolutional networks are still not mature, and there is no effective operation for hypergraph pooling currently. Therefore, a hypergraph pooling network with a self-attention mechanism is proposed. Using a hypergraph structure for data modeling, this model can learn node hidden features with high-order data information through hypergraph convolution operation which introduces a self-attention mechanism, select important nodes both on structure and content through hypergraph pooling operation, and then obtain more accurate hypergraph representation. Experiments on text classification, dish classification, and protein classification tasks show that the proposed method outperforms recent state-of-the-art methods.     

HEp-2 image classification using intensity order pooling based features and bag of words

Human Epithelial type 2 (HEp-2) cells play an important role in the diagnosis of autoimmune disorder. Traditional approach relies on specialists to observe HEp-2 slides via the fluorescence microscope, which suffers from a number of shortcomings like being subjective and labor intensive. Pattern recognition techniques have been recently introduced to this research issue to make the process automatic. However, performances of current systems available in literature are not satisfying. We propose in this paper a framework using intensity order pooling based gradient feature and bag of words for HEp-2 classification. By pooling the gradient features based on the intensity orders of local grid points, the pooled feature is rotationally invariant without requirement of orientation estimation. The proposed approach was fully tested using publicly available ICPR dataset and our own SZU dataset. Experimental results show that the propose method significantly outperformed widely used SIFT feature and the winner of ICPR contest 2012. Encouraging 100% image level accuracy was achieved on the SZU dataset.     

基于ARCNN-GAP网络的语音情感识别

语音情感识别是计算机理解人类情感最直接的方式,是实现人机交互智能化的重要渠道,但识别模型的性能需要进一步提升。为实现这一目标,提出一种基于循环卷积神经网络的语音情感识别模型ARCNN-GAP。其中,循环卷积层具有弹性路径,在确保网络深度的同时能保证优化时的梯度回传,提取更加有效的情感特征;全局平均池化运算可以在减少计算复杂度的同时降低过拟合风险;而注意力机制能够使模型更多关注情感相关特征。使用韵律特征和谱特征的融合特征在CASIA和EMO-DB数据库上进行研究,分别取得了83.29%和75.28%的识别率。实验结果表明：ARCNN-GAP具有更好的识别性能和泛化性。     

Natural disaster risk in Asian megacities: A case for risk pooling?

This paper examines the characteristics of natural disaster risk in the context of megacities generally, and Asian megacities particularly. A key gap in approaches to managing megacity risk has been attention to the financial aspects, for which interest has lately been emerging in terms of exploring whether such risk may be suitable for a donor-assisted regional Asian risk pooling scheme. One suggestion concerns insuring public sector liabilities in terms of infrastructure replacement funding, liquidity support and relief to the population. Recently, this was operationalized in the Caribbean regional pooling of hurricane and earthquake risks, and by the Mexican government for earthquake risks. In both cases, central governments are the actors. We assess the rationale and applicability of such deliberations given the dynamic nature of vulnerability and risk, and discuss conditions for conducting similar transactions for Asian megacity risks. Overall, given our adopted criteria, we tentatively conclude that there may indeed be a case for risk pooling, yet the dynamics of assessing formal and informal risks as well as the specificity of conditions in respective megacities pose important hurdles that have to be overcome.     

The ancient black art and transdisciplinary extent of pseudoreplication.

The history, definitions, and transdisciplinary extent of pseudoreplication, as well as some key concepts of experimental design, are briefly reviewed. Pseudoreplication, sometimes also referred to as the ’unit of analysis error,’ is one of the commonest errors of statistical analysis and interpretation. It is a simple albeit serious one. It persists in part because of the failure of statisticians and scientists to develop a clear, consistent terminology for the concepts of statistics, experimental design, and sampling design that is used across all disciplines, as well as a terminology for specific categories of the more common errors. Lack of a clear terminology, in turn, has fostered narrow, discipline-specific jargon, inconsistency among textbooks and reference works, and ineffective teaching. Reform of terminology is possible, and great improvement in statistical practice would follow. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A management scheme for improving transportation efficiency and contributing to the enhancement of the social fabric

The continuously increasing need for mobility has brought about not only significant facilities in several aspects of human initiative, but also growing traffic congestions, a phenomenon that leads to unpleasant everyday situations at a short time level, but in the long run also to the degradation of the level of quality of living in large cities and the alienation between people. The management of traffic stands, thus, as a fundamental prerequisite for confronting those issues, enhancing transportation and improving the social fabric. This paper considers the concept of car pooling as a structured approach to this problem, by specifying, developing and validating a mobile-community-driven system for collaborative transportation, namely the “transportation management-car pooling system”. This system is capable of proposing optimal, reliable and secure community matches (taking into consideration personality features, talking interests, driving style, etc.), based on user profile and context information. The paper describes the transportation management-car pooling system, presenting its input parameters, decision making process and outcomes. Finally, indicative simulation results showcase its effectiveness.     

文昌A凹陷古近系低渗凝析气藏成藏特征与成藏模式

低渗凝析气藏的成藏机理是影响文昌A凹陷天然气勘探开发成效的关键问题。运用烃源岩地球化学、储层沉积学和成藏综合分析等方法,对天然气地质条件、气藏特征及成藏动态过程进行系统研究。得出如下新认识:①研究区下渐新统恩平组一段厚层湖沼相、滨浅湖相泥岩是主要的优质气源岩,该套烃源岩分别与恩平组二段、珠海组发育的(扇)三角洲相、潮坪相储层形成 “自生自储” 和“下生上储”2套成藏组合;②恩平组、珠海组储层埋深大于3 500m,孔隙度小于12％,渗透率小于1×10^-3μm,沉积微相和成岩作用控制了“甜点”储层的分布;③恩平组、珠海组储层“先致密、后成藏”致使油气的运移距离较短,具备“近源运移、晚期成藏”的特征。根据上述新认识提出2种成藏模式,即陡坡带垂向运聚模式和缓坡带垂向＋侧向运聚模式,并指出北部缓坡带是下一步新领域拓展勘探的重要方向。     

基于全局多粒度池化的可见光红外行人重识别

可见光红外行人重新识别是一种跨模态检索的问题。由于可见光和红外图像模态差异较大,能够精确的匹配行人仍然具有很大的挑战。最近的研究表明,利用池化描述身体部位的局部特征以及人图像本身的全局特征,即使在身体部位缺失的情况下,也能给出鲁棒的特征表示,但是简单的全局平均池化很难获取行人的细节特征。针对这个问题,本文提出一种新的全局多粒度池化的方法,利用全局平均池化和全局最大池化结合的方法,提取行人更多的背景和纹理信息。此外,传统的三元组损失在跨模态行人重识别上效果并不好。我们设计了一种新的跨模态三元损失,以优化类内和类间距离,并监督网络学习有区别的特征表示。本文通过实验证明了所提方法的有效性,并在RegDB和SYSU-MM01数据集上分别取得了88.01%Rank-1,79.26%mAP,和60.24%Rank-1,57.50%mAP的结果。     

中国天然气基础地质理论问题研究新进展

经过30多年的天然气科技攻关研究,已经形成了具有中国特色的天然气成气、成藏地质理论,但随着国内天然气勘探难度的增大和供需矛盾的突显,需要不断发展和完善天然气地质理论,以指导寻找和发现更多的天然气资源。为此,“十一五”期间开展了有关大气田形成基础地质理论问题的攻关研究,取得了一系列重要进展和认识,主要包括：①3种成因类型天然气（煤成气、原油裂解气和生物气）的生成下限值降低;②高温高压条件下（温度超过150 ℃、压力超过40 MPa）砂岩溶蚀速率增大2～3倍;③建立了气藏盖层的多因素综合定量评价方法,盖层评价逐步走向定量化;④研究了5类气藏（大面积低渗透砂岩、礁滩与岩溶、火山岩、超高压、生物气）的形成条件与主控因素,建立了其成藏模式、明确了其分布规律。进而深化了3大领域的天然气勘探理论：①发展了近源高效聚集的低渗透砂岩气藏勘探理论,推动了苏里格气田、川中须家河组气藏天然气储量快速增长;②创立了以生烃断槽为基本单元的火山岩气藏勘探理论,实现了松辽盆地深层中小断陷火山岩气藏勘探的系列突破;③深化了碳酸盐岩台缘礁滩气藏勘探理论,有效指导了龙岗气田主体评价勘探与向西向东扩展以及四川盆地的天然气风险勘探。