A framework for integrating data alignment, distribution, andredistribution in distributed memory multiprocessors

Parallel architectures with physically distributed memory provide a cost-effective scalability to solve many large scale scientific problems. However, these systems are very difficult to program and tune. In these systems, the choice of a good data mapping and parallelization strategy can dramatically improve the efficiency of the resulting program. In this paper, we present a framework for automatic data mapping in the context of distributed memory multiprocessor systems. The framework is based on a new approach that allows the alignment, distribution, and redistribution problems to be solved together using a single graph representation. The Communication Parallelism Graph (CPG) is the structure that holds symbolic information about the potential data movement and parallelism inherent to the whole program. The CPG is then particularized for a given problem size and target system and used to find a minimal cost path through the graph using a general purpose linear 0-1 integer programming solver. The data layout strategy generated is optimal according to our current cost and compilation models     

A fuzzy graph matching approach in intelligence analysis and maintenance of continuous situational awareness

In intelligence analysis a situation of interest is commonly obscured by the more voluminous amount of unimportant data. This data can be broadly divided into two categories, hard or physical sensor data and soft or human observed data. Soft intelligence data is collected by humans through human interaction, or human intelligence (HUMINT). The value and difficulty in manual processing of these observations due to the volume of available data and cognitive limitations of intelligence analysts necessitate an information fusion approach toward their understanding. The data representation utilized in this work is an attributed graphical format. The uncertainties, size and complexity of the connections within this graph make accurate assessments difficult for the intelligence analyst. While this graphical form is easier to consider for an intelligence analyst than disconnected multi-source human and sensor reports, manual traversal for the purpose of obtaining situation awareness and accurately answering priority information requests (PIRs) is still infeasible. To overcome this difficulty an automated stochastic graph matching approach is developed. This approach consists of three main processes: uncertainty alignment, graph matching result initialization and graph matching result maintenance. Uncertainty alignment associates with raw incoming observations a bias adjusted uncertainty representation representing the true value containing spread of the observation. The graph matching initialization step provides template graph to data graph matches for a newly initialized situation of interest (template graph). Finally, the graph matching result maintenance algorithm continuously updates graph matching results as incoming observations augment the cumulative data graph. Throughout these processes the uncertainties present in the original observations and the template to data graph matches are preserved, ultimately providing an indication of the uncertainties present in the current situation assessment. In addition to providing the technical details of this approach, this paper also provides an extensive numerical testing section which indicates a significant performance improvement of the proposed algorithm over a leading commercial solver.     

基于ADAMS的压紧机构运动学仿真研究

作为传统机械结构中常见的典型机构之一,压紧机构广泛运用于不同机械设备中,因此具有一定的研究价值。本文选取研究对象是普通小型压力机,其压紧机构实质是曲柄连杆结构,研究目的是研究该机构的运动规律,为以后研究更加相对复杂的机械结构提供一定的参考价值。首先利用ADAMS建立压力机的虚拟样机,再利用ADAMS完成运动学仿真分析,最后得到了压紧过程的仿真动画和工件所受压力变化的曲线图,曲线图直观反映了压紧机构的运动规律。分析曲线图表明,曲线图上压紧力的变化符合实际情况,为进一步深入研究提供参考。     

Mining periodic movement patterns of mobile phone users based on an efficient sampling approach

In m-commerce services, the periodic movement trends of customers at specific periods can be adopted to allocate the resources of telecommunications systems effectively and offer personalized location-based services. This study explores the mining of periodic maximal promising movement patterns. A detailed process for mining periodic maximal promising movement patterns based on graph mapping and sampling techniques is devised to enhance mining efficiency. First, a random sample of movement paths from time intervals is taken. Second, a unique path graph structure is built to store the movement paths obtained from the sample. Third, a graph traversal algorithm is developed to identify the maximal promising movement patterns. Finally, vector operations are undertaken to examine the maximal promising movement patterns in order to derive the periodic maximal promising movement patterns. Experimental results reveal that the sampling approach with mining has excellent execution efficiency and scalability in the investigation of periodic maximal promising movement patterns.     

一种新的基于频繁闭显露模式的图分类方法

由于图模型能够准确地表示科学与工程领域中数据的关键特征,图挖掘逐渐成为了数据挖掘领域的热点研究内容.图分类是图挖掘的一个重要研究分支.提出了一种新的基于频繁闭显露模式的图分类方法CEP,其基本思想是首先挖掘频繁闭图模式,然后从闭图模式中得到显露模式,最后根据显露模式构造一系列分类规则.实验结果显示:在对化合物数据分类时,CEP在分类性能上优于目前最好的图分类方法.而且,领域专家容易理解和利用CEP产生的分类规则.     

关系敏感型多子图图神经网络的多模态实体对齐

作为融合多源异构知识图谱的主要手段, 实体对齐一般首先编码实体等图结构信息, 而后通过计算实体间相似度来获取对齐实体. 然而, 现存的多模态对齐方法往往直接引入预训练方法表达模态特征, 忽略了模态间的融合以及模态特征与图结构间的融合. 因此, 本文提出一种关系敏感型的多子图图神经网络(RAMS)方法. 通过多子图图神经网络编码方法对模态信息与图结构进行结合并获得实体表征, 通过跨域相似度计算得到对齐结果. 广泛且多角度的实验证明了本文所提出的模型在准确率、效率、鲁棒性方面均超过了基线模型.     

基于最大权值路径算法的DNA多序列比对方法

针对生物序列分析中的多序列比对问题,当输入数据量比较大时,人们提出了很多启发式的算法来改善计算速度和比对结果.提出了用于进行全局DNA多序列比对的一种方法:MWPAlign(maximum weighted path alignment).该算法把序列信息用de Bruijn图的形式表示,并将输入序列的信息记录在图的边上,这样,就将求调和序列的问题转化为求图的最大权值路径问题,使多序列比对问题的时间复杂度降低到几乎线性.实验结果显示:MWPAlign是可行的多序列比对算法,尤其对于变异率低于5.2%的大量序列数据,相对于CLUSTALW(cluster alignments weight),T-Coffee和HMMT(hidden Markov model training)有较好的比对结果和运算性能.     

基于双向对齐与属性信息的跨语言实体对齐

实体对齐表示在不同的知识图谱中查找引用相同现实身份的实体。目前主流的基于图嵌入的实体对齐方法中的对齐实体通常具有相似的属性,有效利用属性信息可提升实体对齐效果,同时由于不同知识图谱之间的知识分布差异,仅考虑单个方向的对齐预测会导致预测结果出现偏差。针对上述问题,提出一种改进的跨语言实体对齐方法。利用融合属性信息的双向对齐图卷积网络模型,将前馈神经网络编码实体对应的属性信息与初始的实体嵌入相结合,得到联合属性信息的实体表示,并使用双向对齐机制实现跨语言的实体对齐预测。在3个跨语言数据集上的实验结果表明,该方法通过融合更多的知识图谱信息增强了实体表示能力,并且利用双向对齐机制缓解了数据分布差异问题,相比基于图嵌入的实体对齐方法整体性能更优。     

A robust projection plane selection strategy for UAV image stitching

Unmanned Aerial Vehicles (UAVs) are the most popular way to collect ground data today, thanks to their low cost and matchless convenience. However, UAVs are prone to unstable flight poses because they are so light in weight, which has resulted in a new challenge for UAV image stitching. In this paper, we propose a robust approach to stitch UAV images captured from approximately planar scenes without pose parameters assistance. The key idea of the proposed framework lies in an effective projection plane selection strategy, which is capable of resisting the perspective distortion from existing pose-perturbed images. To select a reasonable reference image as the projection plane, we first divide all the images into two groups (stable group and unstable group) according to their registration error under the affine model. Then, a specifically designed approach is used to define a weighted topological graph, which guarantees that the reference image is selected from the stable group while maintaining a global minimum accumulated registration error. Based on our cost topological graph, each unstable group image is locally attached to a stable group image via a homography. Finally, alignment parameters of all the stable group images are solved using affine model, after which global optimization is performed on the model of both groups. Comparing our results to those of the conventional approaches indicates that our proposed approach produced superior results in several challenging experiments involving both qualitative and quantitative evaluation.     

Inertial Sensed Ego‐motion for 3D Vision

Inertial sensors attached to a camera can provide valuable data about camera pose and movement. In biological vision systems, inertial cues provided by the vestibular system are fused with vision at an early processing stage. In this article we set a framework for the combination of these two sensing modalities. Cameras can be seen as ray direction measuring devices, and in the case of stereo vision, depth along the ray can also be computed. The ego‐motion can be sensed by the inertial sensors, but there are limitations determined by the sensor noise level. Keeping track of the vertical direction is required, so that gravity acceleration can be compensated for, and provides a valuable spatial reference. Results are shown of stereo depth map alignment using the vertical reference. The depth map points are mapped to a vertically aligned world frame of reference. In order to detect the ground plane, a histogram is performed for the different heights. Taking the ground plane as a reference plane for the acquired maps, the fusion of multiple maps reduces to a 2D translation and rotation problem. The dynamic inertial cues can be used as a first approximation for this transformation, allowing a fast depth map registration method. They also provide an image independent location of the image focus of expansion and center of rotation useful during visual based navigation tasks. © 2004 Wiley Periodicals, Inc.     

基于图表示和匹配的表单定位与提取

为了实现对不同类型、分辨率和方向的快递表单上用户感兴趣区域信息的获取,本文提出了一种基于图表示和匹配的表单定位与提取方法。选择参考表单中已有的印刷图案或字符等关键区域作为基准位置,进行图的表示。基于图像分割得到的候选关键区域对待处理表单进行图表示。然后,根据图的属性计算待处理表单与参考表单的相似度。最后,将最大相似度对应的同构图作为参考表单图的最优匹配,并建立同构图与参考表单图位置映射,定位出表单。本文实验数据集来源于真实场景下采集的快递包裹表单图像。实验结果表明:本文算法在快递包裹表单图像上具有良好的性能,对旋转、光照变化、局部遮挡具有较好的鲁棒性。     

Modeling Massive RFID Data Sets: A Gateway-Based Movement Graph Approach

Massive Radio Frequency Identification (RFID) data sets are expected to become commonplace in supply chain management systems. Warehousing and mining this data is an essential problem with great potential benefits for inventory management, object tracking, and product procurement processes. Since RFID tags can be used to identify each individual item, enormous amounts of location-tracking data are generated. With such data, object movements can be modeled by movement graphs, where nodes correspond to locations and edges record the history of item transitions between locations. In this study, we develop a movement graph model as a compact representation of RFID data sets. Since spatiotemporal as well as item information can be associated with the objects in such a model, the movement graph can be huge, complex, and multidimensional in nature. We show that such a graph can be better organized around gateway nodes, which serve as bridges connecting different regions of the movement graph. A graph-based object movement cube can be constructed by merging and collapsing nodes and edges according to an application-oriented topological structure. Moreover, we propose an efficient cubing algorithm that performs simultaneous aggregation of both spatiotemporal and item dimensions on a partitioned movement graph, guided by such a topological structure.     

Mining sequential patterns from data streams: a centroid approach

In recent years, emerging applications introduced new constraints for data mining methods. These constraints are typical of a new kind of data: the data streams. In data stream processing, memory usage is restricted, new elements are generated continuously and have to be considered in a linear time, no blocking operator can be performed and the data can be examined only once. At this time, only a few methods has been proposed for mining sequential patterns in data streams. We argue that the main reason is the combinatory phenomenon related to sequential pattern mining. In this paper, we propose an algorithm based on sequences alignment for mining approximate sequential patterns in Web usage data streams. To meet the constraint of one scan, a greedy clustering algorithm associated to an alignment method is proposed. We will show that our proposal is able to extract relevant sequences with very low thresholds.     

基于属性嵌入与图注意力网络的实体对齐算法

实体对齐旨在找到位于不同知识图谱中的等效实体,是实现知识融合的重要步骤.当前主流的方法是基于图神经网络的实体对齐方法,这些方法往往过于依赖图的结构信息,导致在特定图结构上训练得到的模型不能拓展应用于其他图结构中.同时,大多数方法未能充分利用辅助信息,例如属性信息.为此,本文提出了一种基于图注意力网络和属性嵌入的实体对齐方法,该方法使用图注意力网络对不同的知识图谱进行编码,引入注意力机制从实体应用到属性,在对齐阶段将结构嵌入和属性嵌入进行结合实现实体对齐效果的提升.在现实世界的3个真实数据集上对本文模型进行了验证,实验结果表明提出的方法在很大程度上优于基准的实体对齐方法.     

基于MFE-BERT与FNNAttention的心理医学知识图谱构建

针对心理医学领域文本段落冗长、数据稀疏、知识散乱且规范性差的问题, 提出一种基于多层级特征抽取能力预训练模型(MFE-BERT)与前向神经网络注意力机制(FNNAttention)的心理医学知识图谱构建方法. MFE-BERT在BERT模型基础上将其内部所有Encoder层特征进行合并输出, 以获取包含更多语义的特征向量, 同时对两复合模型采用FNNAttention机制强化词级关系, 解决长文本段落语义稀释问题. 在自建的心理医学数据集中, 设计MFE-BERT-BiLSTM-FNNAttention-CRF和MFE-BERT-CNN-FNNAttention复合神经网络模型分别进行心理医学实体识别和实体关系抽取, 实体识别F1值达到93.91%, 实体关系抽精确率达到了89.29%, 通过融合文本相似度与语义相似度方法进行实体对齐, 将所整理的数据存储在Neo4j图数据库中, 构建出一个含有3652个实体, 2396条关系的心理医学知识图谱. 实验结果表明, 在MFE-BERT模型与FNNAttention机制的基础上构建心理医学知识图谱切实可行, 提出的改进模型所搭建的心理医学知识图谱可以更好地应用于心理医学信息管理中, 为心理医学数据分析提供参考.     

Recovery of missing information in graph sequences by means of reference pattern matching and decision tree learning

Algorithms for the analysis of graph sequences are proposed in this paper. In particular, we study the problem of recovering missing information and predicting the occurrence of nodes and edges in time series of graphs. Two different recovery schemes are developed. The first scheme uses reference patterns that are extracted from a training set of graph sequences, while the second method is based on decision tree induction. Our work is motivated by applications in computer network analysis. However, the proposed recovery and prediction schemes are generic and can be applied in other domains as well.     

方面级多模态协同注意图卷积情感分析模型

目的 方面级多模态情感分析日益受到关注,其目的是预测多模态数据中所提及的特定方面的情感极性。然而目前的相关方法大都对方面词在上下文建模、模态间细粒度对齐的指向性作用考虑不够,限制了方面级多模态情感分析的性能。为了解决上述问题,提出一个方面级多模态协同注意图卷积情感分析模型（aspect-level multimodal co-attention graph convolutional sentiment analysis model,AMCGC）来同时建模方面指向的模态内上下文语义关联和跨模态的细粒度对齐,以提升情感分析性能。方法 AMCGC为了获得方面导向的模态内的局部语义相关性,利用正交约束的自注意力机制生成各个模态的语义图。然后,通过图卷积获得含有方面词的文本语义图表示和融入方面词的视觉语义图表示,并设计两个不同方向的门控局部跨模态交互机制递进地实现文本语义图表示和视觉语义图表示的细粒度跨模态关联互对齐,从而降低模态间的异构鸿沟。最后,设计方面掩码来选用各模态图表示中方面节点特征作为情感表征,并引入跨模态损失降低异质方面特征的差异。结果 在两个多模态数据集上与9种方法进行对比,在Twitter-2015数据集中,相比于性能第2的模型,准确率提高了1.76%;在Twitter-2017数据集中,相比于性能第2的模型,准确率提高了1.19%。在消融实验部分则从正交约束、跨模态损失、交叉协同多模态融合分别进行评估,验证了AMCGC模型各部分的合理性。结论 本文提出的AMCGC模型能更好地捕捉模态内的局部语义相关性和模态之间的细粒度对齐,提升方面级多模态情感分析的准确性。