基于随机游走的网络表示学习推荐算法

根据网络结构中的连接关系得到节点的向量表示,进而将节点的向量表示应用于推荐算法可有效提升其建模能力。针对推荐系统中的同质网络,提出结合随机游走的网络表示学习推荐算法。以DeepWalk算法为基础,在随机游走过程中根据节点重要性设定节点游走序列数,并设置终止概率以控制游走长度优化采样结果,在网络表示学习过程中将SkipGram模型融合节点属性信息,同时考虑上下文节点离中心节点的距离获得更准确的推荐结果。实验结果表明,该算法相比DeepWalk、Node2vec等算法具有更高的推荐准确度,并且较好地解决了冷启动问题。     

基于网络表示学习与随机游走的链路预测算法

现有的基于随机游走链路预测指标在无权网络上的转移过程存在较强随机性,没有考虑在网络结构上不同邻居节点间的相似性对转移概率的作用。针对此问题,提出一种基于网络表示学习与随机游走的链路预测算法。首先,通过基于深度学习的网络表示学习算法——DeepWalk学习网络节点的潜在结构特征,将网络中的各节点表征到低维向量空间;然后,在重启随机游走（RWR）和局部随机游走（LRW）算法的随机游走过程中融合各邻居节点在向量空间上的相似性,重新定义出邻居节点间的转移概率;最后,在5个真实数据集上进行大量实验验证。实验结果表明：相比8种具有代表性的基于网络结构的链路预测基准算法,所提算法链路预测结果的AUC值均有提升,最高达3.34%。     

一种随机游走中心性的快速算法

自然网络都具有一定的聚簇结构, 聚簇之间的节点称之为桥节点, 桥节点对网络的流通性有着重要的作用。发现桥节点, 能够找到网络最为脆弱的部分。在随机游走中心性的基础上, 提出一种计算网络桥节点的快速算法。通过人工合成以及在自然网络上进行实验, 结果表明算法能够很好地发现各种网络的桥节点。     

基于改进随机游走的网络表示学习算法

现有的基于Word2vec的网络表示学习（NRL）算法使用随机游走（RW）来生成节点序列,针对随机游走倾向于选择具有较大度的节点,生成的节点序列不能很好地反映网络结构信息,从而影响表示学习性能的问题,提出了基于改进随机游走的网络表示学习算法。首先,使用RLP-MHRW算法生成节点序列,它在生成节点序列时不会偏向大度节点,得到的节点序列能更好地反映网络结构信息;然后,将节点序列投入到Skip-gram模型得到节点表示向量;最后,利用链路预测任务来测度表示学习性能。在4个真实网络数据集上进行了实验。在论文合作网络arXiv ASTRO-PH上与LINE和node2vec算法相比,链路预测的AUC值分别提升了8.9%和3.5%,其他数据集上也均有提升。实验结果表明,RLP-MHRW能有效提高基于Word2vec的网络表示学习算法的性能。     

一种改进的随机游走图像分割算法

为克服传统以像素为单位的随机游走算法随着像素点增多,计算量大、复杂度高、分割速度大大减慢,且对彩色图像分割效果不理想的问题,提出一种改进的随机游走图像分割方法。首先,使用改进和优化后的分水岭算法对目标图像进行预分割,为防止分水岭算法过分割问题,结合使用非线性各向异性扩散方法和形态学处理方法进行处理;然后,将分水岭算法分割后形成的同质区域作为图的节点用于随机游走算法,通过用户标记种子区域,分割出感兴趣的目标物体;最后,给出传统随机游走分割方法与本文提出的分割方法的实验结果比较,并对它们进行分析和评价。     

一种基于随机游走的歌曲推荐算法

推荐系统可以帮助人们在海量的数据中发现所需的有价值的信息。传统的协同过滤推荐算法根据历史数据中用户对项目的各种行为操作构建用户-项目评分矩阵,进而计算相似度,从而预测用户对项目的偏好程度进行推荐。但因为评分数据通常较为稀疏,使得推荐的准确性不高,从而不能很好地对用户进行推荐。针对这个问题,提出一种结合场论理论的随机游走歌曲推荐算法,融合歌曲评分相似度和歌曲基本信息相似度,降低歌曲间综合相似度矩阵的稀疏性,并将物理学中的场论理论和歌曲的重要度结合,构造转移概率矩阵,从而实现歌曲推荐。实验表明,该算法较协同过滤算法的推荐准确性更佳。     

一种融合差分隐私的随机游走算法

如今的信息化时代,用户之间的社交网络信息越发详细,发布这些网络数据经常会威胁到一些个人隐私.而推荐算法中,根据用户物品之间的二分图关系,进行随机游走推荐,能更可靠地推荐目标用户可能选择的物品.由于随机游走复杂度过高,一般将图转化成转移矩阵进行计算,但是游走时无法保证该目标用户以及其他用户的隐私信息.在隐私得不到保护的前...     

一种基于神经网络模型的句子排序方法

句子排序是多文本摘要中的重要问题,合理地对句子进行排序对于摘要的可读性和连贯性具有重要意义。该文首先利用神经网络模型融合了五种前人已经提出过的标准来决定任意两个句子之间的连接强度,这五种标准分别是时间、概率、主题相似性、预设以及继承。其次,该文提出了一种基于马尔科夫随机游走模型的句子排序方法,该方法利用所有句子之间的连接强度共同决定句子的最终排序。最终,该文同时使用人工和半自动方法对句子排序的质量进行评价,实验结果表明该文所提出方法的句子排序质量与基准算法相比具有明显提高。
     

一种基于随机游走模型的多标签分类算法

在数据挖掘领域,传统的单分类和多分类问题已经得到了广泛的研究.但是多标签数据的普遍存在性和重要性直到近些年来才逐渐得到人们的关注.在多标签分类问题中,由于标签相关性的存在,传统的单分类和多分类问题的解决方法,无法简单地应用于多标签分类问题.文中提出了一种基于随机游走模型的多标签分类算法,称为多标签随机游走算法.首先,将多标签数据映射成为多标签随机游走图.当输入一个未分类数据时,建立一个多标签随机游走图系列.而后,对图系列中的每个图应用随机游走模型,得到遍历每个顶点的概率分布,并将这个点概率分布转化成每个标签的概率分布.最后,基于多标签随机游走算法,文中给出了一种新的阈值学习算法.真实数据集上的实验表明,多标签随机游走算法可以有效地解决多标签分类问题.     

A new and simple proof of the equivalence theorem for behaviors

Two full row rank representations of the same behavior are related through a left unimodular transformation. We present a new and extremely simple and insightful proof for this well-established fact.     

先秦汉语排比句自动识别研究--以《孟子》《论语》中的排比句自动识别为例

《论语》、《孟子》等先秦汉语文献对后世影响深远,其中大量使用的排比句,使这些文献有极强的说服力,增强了表达效果。以《孟子》、《论语》中的排比句的自动识别为例,探讨了古代汉语修辞格的自动识别。详细分析先秦汉语排比句的特点,并设计了相应的自动识别算法,来分别处理《孟子》和《论语》中的排比句,并对自动识别结果进行了分析和比较。在处理《孟子》、《论语》这两种风格不太相同的先秦文献时,排比句自动识别算法都保持了较好的性能。     

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.     

A grey wolf optimizer using Gaussian estimation of distribution and its application in the multi-UAV multi-target urban tracking problem

To overcome premature convergence in the grey wolf optimizer (GWO), in this study, a modified GWO integrating the basic GWO with the Gaussian estimation of distribution (GED) strategy, called GEDGWO, is proposed. GEDGWO employs the Gauss probability model to estimate the distribution of the selected superior individuals and shifts the weighted mean to adjust the search directions. Additionally, a Gaussian distribution based inferior solutions repair (ISR) method is introduced to modify the ill-shaped distribution of the population. A disturbed Gaussian random walk method is utilized to strengthen the local exploration ability. The performance of GEDGWO is compared with those of other promising GWO variants and state-of-the-art algorithms on a benchmarking CEC 2014 test suite. Non-parametric Wilcoxon and Friedman tests as well as the post hoc Iman–Davenport test are performed to further verify the efficacy of GEDGWO. Moreover, GEDGWO is applied to solve multi-UAV multi-target urban tracking path planning problem. To overcome the shortcomings of the previous solution model, a new model is described to address this complex real-time engineering optimization problem. The validity and practicability of the problem models as well as the accuracy and efficiency of GEDGWO are demonstrated by the experimental results.     

The interplay between learning and the use of ICT in Rwandan student teachers' everyday practice

The paper describes a study conducted in Rwanda involving 12 participants selected from a larger cohort of 24 final-year university students who were part of a group-based training programme. The programme was about how to search, retrieve, and use web-based literature. Empirical data were collected through interviews and focus group discussions. The purpose was to explore ways of using information and communication technology (ICT) in student teachers' everyday learning practice. The study draws from a sociocultural perspective and emphasis is put on a literature review involving ICT in teacher education. The findings reveal that utilization of ICT pertains to three major types of variation among student teachers who use ICT: passive, reluctant, and active users. The active ICT users demonstrated a capacity to cross group boundaries and play a central role as agents of change in learning practice. The point is that more experienced student teachers can assist their colleagues in the zone of proximal development and, therefore, enhance the integration of the new technology in teacher education. This implies that having access to ICT together with some instruction is not sufficient to prompt students to start using this technology as a pedagogical tool. Moreover, confrontation of different experiences regarding the use of ICT can spearhead change in student teachers' learning practice through critical reflection.     

A new method for analysis and design of iterative learning control algorithms in the time-domain

In this paper, a novel analysis method for iterative learning control (ILC) algorithms is presented. Even though expressed in the lifted system representation and hence in the time-domain, the convergence rate as a function of the frequency content of the error signal can be determined. Subsequently, based on the analysis method, a novel ILC algorithm (F-ILC) is proposed. The convergence rate at specific frequencies can be set directly in the design process, which allows simple tuning and a priori known convergence rates. Using the F-ILC design, it is shown how to predict the required number of iterations until convergence is achieved, depending on the reference trajectory and information on the system repeatability. Numerical examples are given and experimental results obtained on an internal combustion engine test bench are shown for validation.     

A dynamic analysis of the interplay between asynchronous and synchronous communication in online learning: The impact of motivation

With the increased affordances of synchronous communication tools, more opportunities for online learning to resemble face‐to‐face settings have recently become available. However, synchronous communication does not afford as much time for reflection as asynchronous communication. Therefore, a combination of synchronous and asynchronous communication in e‐learning would seem desirable to optimally support learner engagement and the quality of student learning. It is still an open question though, how to best design online learning with a blend of synchronous and asynchronous communication opportunities over time. Few studies have investigated the relationship between learners' actual use of synchronous and asynchronous communication over time. Therefore, this study addressed that relationship in an online course (N = 110), taking into account student motivation, and employing a dynamic inter‐temporal perspective. In line with our assumptions, we found some support for the expected association between autonomous motivation and engagement in asynchronous and synchronous communication, be it restricted primarily to the first course period. Also, positive relations between engagement in synchronous and asynchronous communication were found, with the strongest influence from using asynchronous to synchronous communication. This study adds to the knowledge base needed to develop guidelines on how synchronous communication can be combined with asynchronous learning.     

A review of research on bridging the gap between formal and informal learning with technology in primary school contexts

This study reviews empirical research articles published in the field of technology‐enhanced learning in the out‐of‐class contexts in primary schools between the years 2007 and 2016 and explores how the body of research has connected formal and informal learning experiences, referred to in the paper as bridging the gap. The review focuses on 43 selected experiments from 41 research papers, which are in detail examined and classified using the 3 criteria: (a) the Bloom's taxonomy for learning, (b) the intentionality to physical settings classification used to differentiate types of learning in various environments, and (c) the characteristics of seamless learning design. The findings confirm that technology can enhance learning in and out of classroom, especially by impacting student interest, motivation, and engagement. The close examination of the subset of studies with cognitive gains shows that they successfully bridged the gap between learning spaces and that such bridging positively correlates with the number of steps in the learning activity design. The successful bridging of the gap between learning spaces could further benefit from including more online social learning activities into the designed learning process and from involving teachers as cocreators of the learning process and resources.