一种基于全局和局部标记相关性的多标记分类算法

多标记学习用于处理一个示例同时与多个类别标记相关的问题.在多标记学习中,标记相关性能够显著提升学习算法的性能.大多数现有的多标记学习算法在利用标记的相关性时,要么只使用被所有示例所共享的全局标记相关性,要么就使用局部标记相关性,它们认为不同簇中的示例应该存在不同的标记相关性.本文中,我们提出了一种同时利用全局和局部标记相关性的多标记学习算法,从而为学习进程提供更全面的标记信息.在计算全局和局部标记相关性时,我们使用了余弦相似性来获取不同标记之间的正相关性和负相关性,这样有助于我们进一步实现更可靠的多标记学习.我们在多种类型的数据集上进行了广泛的对比实验来验证所提算法的有效性.实验结果表明,该算法显著优于大多数对比算法,展现出其在多标记学习中的突出性能.     

候选标记信息感知的偏标记学习算法

在偏标记学习中,示例的真实标记隐藏在由一组候选标记组成的标记集中。现有的偏标记学习算法在衡量示例之间的相似度时,只基于示例的特征进行计算,缺乏对候选标记集信息的利用。该文提出一种候选标记感知的偏标记学习算法(CLAPLL),在构建图的阶段有效地结合候选标记集信息来衡量示例之间的相似度。首先,基于杰卡德距离和线性重构,计算出各个示例的标记集之间的相似度,然后结合示例相似度和标记集的相似度构建相似度图,并通过现有的基于图的偏标记学习算法进行学习和预测。3个合成数据集和6个真实数据集上实验结果表明,该文方法相比于基线算法消歧准确率提升了0.3%～16.5%,分类准确率提升了0.2%～2.8%。     

哈希快速多标记学习算法

本文针对多标记学习耗时大、很难处理大规模数据的问题,提出了一种哈希快速多标记学习算法（HFMLL）,该算法将哈希算法与多标记学习算法结合,采用局部敏感哈希算法快速获得每个样本的近邻样本,并通过最小独立置换的MinHash算法快速找到每个标记的相关标记,根据其近邻样本及相关标记的信息,运用最大后验概率准则来预测新样本的标记集。实验表明HFMLL 算法在保持较高分类性能的情况下,算法速度明显优于目前的多标记算法,可以广泛应用于大规模的数据集。      

Flow aggregated, traffic driven label mapping in label-switchingnetworks

Label-switching technology enables high performance and flexible layer-3 packet forwarding based on the fixed-length label information that is mapped to the layer-3 packet stream. A label-switching router (LSR) forwards layer-3 packets based on their layer-3 address information or their label information that is mapped to the layer-3 address information. Two label-mapping policies have been proposed. One is traffic driven mapping, where the label is mapped for a layer-3 packet stream of each host-pair according to the actual packet arrival. The other is topology driven mapping, where the label is mapped in advance for a layer-3 packet stream toward the same destination network, regardless of actual packet arrival to the LSR. This paper evaluates the required number of labels under each of these two label-mapping policies using real backbone traffic traces. The evaluation shows that both label-mapping policies require a large number of labels. In order to reduce the required number of labels, we propose a label-mapping policy that is a combination of the two label-mapping policies above. This is traffic-driven label mapping for the packet stream toward the same destination network. The evaluation shows that the proposed label-mapping policy requires only about one-tenth as many labels as the traffic-driven label mapping for the host-pair packet stream and the topology-driven label mapping for the destination-network packet stream     

A run-based two-scan labeling algorithm.

We present an efficient run-based two-scan algorithm for labeling connected components in a binary image. Unlike conventional label-equivalence-based algorithms, which resolve label equivalences between provisional labels, our algorithm resolves label equivalences between provisional label sets. At any time, all provisional labels that are assigned to a connected component are combined in a set, and the smallest label is used as the representative label. The corresponding relation of a provisional label and its representative label is recorded in a table. Whenever different connected components are found to be connected, all provisional label sets concerned with these connected components are merged together, and the smallest provisional label is taken as the representative label. When the first scan is finished, all provisional labels that were assigned to each connected component in the given image will have a unique representative label. During the second scan, we need only to replace each provisional label by its representative label. Experimental results on various types of images demonstrate that our algorithm outperforms all conventional labeling algorithms.     

基于n元词组表示的去噪方法及其在跨语言映射中的应用

具有结构化输出的学习任务（结构化学习）在自然语言处理领域广泛存在。近年来研究人员们从理论上证明了数据标记的噪声对于结构化学习的巨大影响，因此为适应结构化学习任务的去噪算法提出了需求。受到近年来表示学习发展的启发，本文提出将自然语言的子结构低维表示引入结构化学习任务的样本去噪算法中。这一新的去噪算法通过n元词组的表示为序列标注问题中每个节点寻找近邻，并根据节点标记与其近邻标记的一致性实现去噪。本文在命名实体识别和词性标注任务的跨语言映射上对上述去噪方法进行了验证，证明了这一方法的有效性。     

Label swapping versus label preserving in packet switched networks: Impact on the label space size

In recent years, switching and networking solutions exploiting all-optical nodes are gaining increasing interest to achieve the target of ultrawide-bandwidth and low-latency packet or burst processing. On the one hand, many prototypes, validated by experimental demonstration of all-optical label processing solutions, have been developed. On the other hand, the primitive available technology for performing label processing poses several constraints on the label structures; this in turn significantly impacts the traffic engineering aspects of such a network.In this paper, the label assignment problem is studied in a network that makes forwarding decision based on optical packet labels and formulated independently of any technology. Specifically, the problem of assigning labels to identify the label switched path (LSP) packets in a unique and disjoint way is defined, with the objective of optimally minimizing the label space size (i.e. number of labels, or bits, required to uniquely identify the LSPs). The network scenarios where (i) labels have local point-to-point significance (i.e. the label is swapped when traversing each node), and (ii) labels have end-to-end significance (i.e. the label is preserved along the LSP traversing multiple nodes) are both investigated. For both scenarios, labels can be uniquely identified at each node or at each node-port. The label assignment strategies for all the possible scenarios are investigated.Both theoretical and practical methods, i.e. integer linear programming formulations and heuristics, respectively, are used to assess the efficiency of the proposed label preserving solution. Numerical results show that, for a significant set of network topologies, the label space size increase experienced by networks with label preserving capabilities is limited or negligible in both per-node and per-port label identification.     

Graph-based semi-supervised learning with multiple labels

Conventional graph-based semi-supervised learning methods predominantly focus on single label problem. However, it is more popular in real-world applications that an example is associated with multiple labels simultaneously. In this paper, we propose a novel graph-based learning framework in the setting of semi-supervised learning with multiple labels. This framework is characterized by simultaneously exploiting the inherent correlations among multiple labels and the label consistency over the graph. Based on the proposed framework, we further develop two novel graph-based algorithms. We apply the proposed methods to video concept detection over TRECVID 2006 corpus and report superior performance compared to the state-of-the-art graph-based approaches and the representative semi-supervised multi-label learning methods.     

基于单分类支持向量机和主动学习的网络异常检测研究

对基于支持向量机和主动学习的异常检测方法进行了研究,首先利用原始数据采用无监督方式建立单分类支持向量机模型,然后结合主动学习找出对提高异常检测性能最有价值的样本进行人工标记,利用标记数据和无标记数据以半监督方式对基于单分类支持向量机的异常检测模型进行扩展。实验结果表明,所提方法能够利用少量标记数据获取性能提升,并能够通过主动学习减小人工标记代价,更适用于实际网络环境。     

Frugal incentive mechanism in periodic mobile crowdsensing

Nowadays, numerous incentive mechanisms of mobile crowdsensing have been designed to attract extensive user participation, but most of these mechanisms focus only on independent task scenarios, where the sensing tasks are independent of each other. On the contrary, we focus on a periodical task scenario, where each user participates in the same type of sensing tasks periodically. In this paper, we consider the long‐term user participation incentive in a general periodical mobile crowdsensing system from a frugality payment perspective. We explore the issue under both semi‐online (the intraperiod interactive process is synchronous while the interperiod interactive process is sequential and asynchronous during each period) and online user arrival models (the previous 2 interactive processes are sequential and asynchronous). In particular, we first propose a semi‐online frugal incentive mechanism by introducing a Lyapunov method. Moreover, we also extend it to an online frugal incentive mechanism, which satisfies the long‐term participation constraint and approximate optimality. Finally, extensive simulations show that our mechanisms satisfy the above theoretical properties.     

Repeated communication and Ramsey graphs

We study the savings afforded by repeated use in two zero-error communication problems. We show that for some random sources, communicating one instance requires arbitrarily many bits, but communicating multiple instances requires roughly 1 bit per instance. We also exhibit sources where the number of bits required for a single instance is comparable to the source's size, but two instances require only a logarithmic number of additional bits. We relate this problem to that of communicating information over a channel. Known results imply that some channels can communicate exponentially more bits in two uses than they can in one use     

Robust statistical fusion of image labels

Image labeling and parcellation (i.e., assigning structure to a collection of voxels) are critical tasks for the assessment of volumetric and morphometric features in medical imaging data. The process of image labeling is inherently error prone as images are corrupted by noise and artifacts. Even expert interpretations are subject to subjectivity and the precision of the individual raters. Hence, all labels must be considered imperfect with some degree of inherent variability. One may seek multiple independent assessments to both reduce this variability and quantify the degree of uncertainty. Existing techniques have exploited maximum a posteriori statistics to combine data from multiple raters and simultaneously estimate rater reliabilities. Although quite successful, wide-scale application has been hampered by unstable estimation with practical datasets, for example, with label sets with small or thin objects to be labeled or with partial or limited datasets. As well, these approaches have required each rater to generate a complete dataset, which is often impossible given both human foibles and the typical turnover rate of raters in a research or clinical environment. Herein, we propose a robust approach to improve estimation performance with small anatomical structures, allow for missing data, account for repeated label sets, and utilize training/catch trial data. With this approach, numerous raters can label small, overlapping portions of a large dataset, and rater heterogeneity can be robustly controlled while simultaneously estimating a single, reliable label set and characterizing uncertainty. The proposed approach enables many individuals to collaborate in the construction of large datasets for labeling tasks (e.g., human parallel processing) and reduces the otherwise detrimental impact of rater unavailability.     

Web and personal image annotation by mining label correlation with relaxed visual graph embedding

The number of digital images rapidly increases, and it becomes an important challenge to organize these resources effectively. As a way to facilitate image categorization and retrieval, automatic image annotation has received much research attention. Considering that there are a great number of unlabeled images available, it is beneficial to develop an effective mechanism to leverage unlabeled images for large-scale image annotation. Meanwhile, a single image is usually associated with multiple labels, which are inherently correlated to each other. A straightforward method of image annotation is to decompose the problem into multiple independent single-label problems, but this ignores the underlying correlations among different labels. In this paper, we propose a new inductive algorithm for image annotation by integrating label correlation mining and visual similarity mining into a joint framework. We first construct a graph model according to image visual features. A multilabel classifier is then trained by simultaneously uncovering the shared structure common to different labels and the visual graph embedded label prediction matrix for image annotation. We show that the globally optimal solution of the proposed framework can be obtained by performing generalized eigen-decomposition. We apply the proposed framework to both web image annotation and personal album labeling using the NUS-WIDE, MSRA MM 2.0, and Kodak image data sets, and the AUC evaluation metric. Extensive experiments on large-scale image databases collected from the web and personal album show that the proposed algorithm is capable of utilizing both labeled and unlabeled data for image annotation and outperforms other algorithms.     

基于纹理特征的多区域水平集图像分割方法

现有多区域水平集方法大多利用复杂的能量函数来驱动多个水平集函数的演变,这样不仅模型复杂且存在很多限制.为此本文提出一种基于纹理特征的多区域水平集方法,利用任意数量的水平集函数来对相应数量的图像区域进行分割.本文首先对图像的颜色和纹理信息建立联合分布并将其代入能量函数;引入平滑概率标签,根据概率性质建立基于标签驱动的多区域水平集迭代更新方程.之后将每个水平集投影到离散概率空间得到一系列近似标签,并由这些标签得到基于多区域水平集的先验概率,从而将多个轮廓演变信息代入统计框架.而不同区域的统计参数也通过最小化能量函数由概率标签迭代更新.通过与其他分割算法在大量复杂实景图像上的实验对比,验证了本文算法的有效性.     

基于标记密度分类间隔面的组类属属性学习

类属属性学习避免相同属性预测全部标记,是一种提取各标记独有属性进行分类的一种框架,在多标记学习中得到广泛的应用。而针对标记维度较大、标记分布密度不平衡等问题,已有的基于类属属性的多标记学习算法普遍时间消耗大、分类精度低。为提高多标记分类性能,该文提出一种基于标记密度分类间隔面的组类属属性学习(GLSFL-LDCM)方法。首先,使用余弦相似度构建标记相关性矩阵,通过谱聚类将标记分组以提取各标记组的类属属性,减少计算全部标记类属属性的时间消耗。然后,计算各标记密度以更新标记空间矩阵,将标记密度信息加入原标记中,扩大正负标记的间隔,通过标记密度分类间隔面的方法有效解决标记分布密度不平衡问题。最后,通过将组类属属性和标记密度矩阵输入极限学习机以得到最终分类模型。对比实验充分验证了该文所提算法的可行性与稳定性。

     

A novel semi-supervised canonical correlation analysis and extensions for multi-view dimensionality reduction

Canonical correlation analysis (CCA) is an efficient method for dimensionality reduction on two-view data. However, as an unsupervised learning method, CCA cannot utilize partly given label information in multi-view semi-supervised scenarios. In this paper, we propose a novel two-view semi-supervised learning method, called semi-supervised canonical correlation analysis based on label propagation (LPbSCCA). LPbSCCA incorporates a new sparse representation based label propagation algorithm to infer label information for unlabeled data. Specifically, it firstly constructs dictionaries consisting of all labeled samples; and then obtains reconstruction coefficients of unlabeled samples using sparse representation technique; at last, by combining given labels of labeled samples, estimates label information for unlabeled ones. After that, it constructs soft label matrices of all samples and probabilistic within-class scatter matrices in each view. Finally, in order to enhance discriminative power of features, it is formulated to maximize the correlations between samples of the same class from cross views, while minimizing within-class variations in the low-dimensional feature space of each view simultaneously. Furthermore, we also extend a general model called LPbSMCCA to handle data from multiple (more than two) views. Extensive experimental results from several well-known datasets demonstrate that the proposed methods can achieve better recognition performances and robustness than existing related methods.     

Ubiquitously Supervised Subspace Learning

In this paper, our contributions to the subspace learning problem are two-fold. We first justify that most popular subspace learning algorithms, unsupervised or supervised, can be unitedly explained as instances of a ubiquitously supervised prototype. They all essentially minimize the intraclass compactness and at the same time maximize the interclass separability, yet with specialized labeling approaches, such as ground truth, self-labeling, neighborhood propagation, and local subspace approximation. Then, enlightened by this ubiquitously supervised philosophy, we present two categories of novel algorithms for subspace learning, namely, misalignment-robust and semi-supervised subspace learning. The first category is tailored to computer vision applications for improving algorithmic robustness to image misalignments, including image translation, rotation and scaling. The second category naturally integrates the label information from both ground truth and other approaches for unsupervised algorithms. Extensive face recognition experiments on the CMU PIE and FRGC ver1.0 databases demonstrate that the misalignment-robust version algorithms consistently bring encouraging accuracy improvements over the counterparts without considering image misalignments, and also show the advantages of semi-supervised subspace learning over only supervised or unsupervised scheme.     

用于大规模深度卷积分类网络DOA估计的标签分解方法

为了有效地解决使用深度神经网络求解波达方向(DOA)估计涉及到的大规模分类器的训练和部署实现,本文提出将传统的one-hot分类器分解为多个类别互质的小分类器,然后联合使用多个互质分类器的分类结果重构原始one-hot标签.首先使用标签分解,将原始标签分解为多个互质的小标签,小标签对应的类别为原始类别对质数取余数的结果...     

使用区间路径处理XML查询

近年来,在XML查询处理方法中发表了一些基于节点流栈连接的高效的分枝连接算法。然而,这些算法普遍存在这样的问题:由于它们必须扫描查询中出现的每一个元素对应的节点流,当XML节点数量很大时,查询处理的输入代价很大,效率变得低下。为了解决这个问题,提出了一个新型的标记法记为区间路径,不同于节点流的区间标记法,区间路径可以把具有相同路径的节点集索引到一个集合中。继而提出了分枝点连接算法用于XML查询处理。同基于节点流栈的分枝连接算法相比,该算法有以下优势:节点集的祖先信息直接位于区间路径中;只有和查询结果相关的节点集会被扫描到,大大降低了输入代价;支持查询通配符;对于类型为根路径的查询,只需一次输入操作代价完成查询处理。实验结果表面该算法在输入代价,执行时间和延展性方面都优于基于节点流的分枝连接算法。     

A new label-based source routing for multi-ring networks

This paper presents a new source routing technique for ring and multi-ring networks, which uses short address labels. The main objectives for having this new method are that in case of one or more failures a frame will be guaranteed: (1) to be removed from the ring-termination property, and (2) to be copied at most once and only by its destinations-safety property. The scheme is based on dividing the label address space of each ring into subspaces, such that the address subspaces are physically disjoint. More specifically, each ring, in a multi-ring network, is divided into two or more parts such that adjacent address subspaces are disjoint. The route of each frame is described by a sequence of short address labels in the frame's header. The current route of a frame is determined by the first address label in its header, and it can be used for routing over at most one subspace of the ring