Joint modelling of voicing label and continuous F0 for HMM based speech synthesis

Fundamental frequency, or F0 is critical for high quality speech synthesis in HMM based speech synthesis. Traditionally, F0 values are considered to depend on a binary voicing decision such that they are continuous in voiced regions and undefined in unvoiced regions. Multi-space distribution HMM (MSDHMM) has been used for modelling the discontinuous F0. Recently, a continuous F0 modelling framework has been proposed and shown to be effective, where continuous F0 observations are assumed to always exist and voicing labels are explicitly modelled by an independent stream. In this paper, a refined continuous F0 modelling approach is proposed. Here, F0 values are assumed to be dependent on voicing labels and both are jointly modelled in a single stream. Due to the enforced dependency, the new method can effectively reduce the voicing classification error. Subjective listening tests also demonstrate that the new approach can yield significant improvements on the naturalness of the synthesised speech. A dynamic random unvoiced F0 generation method is also investigated. Experiments show that it has significant effect on the quality of synthesised speech. © 2011 IEEE.

Semi-supervised video segmentation using tree structured graphical models

We present a novel, implementation friendly and occlusion aware semi-supervised video segmentation algorithm using tree structured graphical models, which delivers pixel labels alongwith their uncertainty estimates. Our motivation to employ supervision is to tackle a task-specific segmentation problem where the semantic objects are pre-defined by the user. The video model we propose for this problem is based on a tree structured approximation of a patch based undirected mixture model, which includes a novel time-series and a soft label Random Forest classifier participating in a feedback mechanism. We demonstrate the efficacy of our model in cutting out foreground objects and multi-class segmentation problems in lengthy and complex road scene sequences. Our results have wide applicability, including harvesting labelled video data for training discriminative models, shape/pose/articulation learning and large scale statistical analysis to develop priors for video segmentation. © 2011 IEEE.

Augmented attribute representations

We propose a new learning method to infer a mid-level feature representation that combines the advantage of semantic attribute representations with the higher expressive power of non-semantic features. The idea lies in augmenting an existing attribute-based representation with additional dimensions for which an autoencoder model is coupled with a large-margin principle. This construction allows a smooth transition between the zero-shot regime with no training example, the unsupervised regime with training examples but without class labels, and the supervised regime with training examples and with class labels. The resulting optimization problem can be solved efficiently, because several of the necessity steps have closed-form solutions. Through extensive experiments we show that the augmented representation achieves better results in terms of object categorization accuracy than the semantic representation alone. © 2012 Springer-Verlag.

Multitask learning without label correspondences

We propose an algorithm to perform multitask learning where each task has potentially distinct label sets and label correspondences are not readily available. This is in contrast with existing methods which either assume that the label sets shared by different tasks are the same or that there exists a label mapping oracle. Our method directly maximizes the mutual information among the labels, and we show that the resulting objective function can be efficiently optimized using existing algorithms. Our proposed approach has a direct application for data integration with different label spaces, such as integrating Yahoo! and DMOZ web directories.

MoT - Mixture of trees probabilistic graphical model for video segmentation

We present a novel mixture of trees (MoT) graphical model for video segmentation. Each component in this mixture represents a tree structured temporal linkage between super-pixels from the first to the last frame of a video sequence. Our time-series model explicitly captures the uncertainty in temporal linkage between adjacent frames which improves segmentation accuracy. We provide a variational inference scheme for this model to estimate super-pixel labels and their confidences in nearly realtime. The efficacy of our approach is demonstrated via quantitative comparisons on the challenging SegTrack joint segmentation and tracking dataset [23].

Classification using log Gaussian Cox processes

McCullagh and Yang (2006) suggest a family of classification algorithms based on Cox processes. We further investigate the log Gaussian variant which has a number of appealing properties. Conditioned on the covariates, the distribution over labels is given by a type of conditional Markov random field. In the supervised case, computation of the predictive probability of a single test point scales linearly with the number of training points and the multiclass generalization is straightforward. We show new links between the supervised method and classical nonparametric methods. We give a detailed analysis of the pairwise graph representable Markov random field, which we use to extend the model to semi-supervised learning problems, and propose an inference method based on graph min-cuts. We give the first experimental analysis on supervised and semi-supervised datasets and show good empirical performance.

基于Gabor小波特征确认多媒体消息数字水印版权标识

为解决基于数字水印的无线多媒体消息版权管理系统对提取后水印标识的自动识别问题,在充分考虑多媒体消息在传播中可能遭受攻击的基础上,提出一种基于Gabor小波特征的标识确认方案.该方案利用这类小波函数确定的滤波器适合局部分析和多方向多尺度分析的特点,提取与水印版权标识结构信息相关的统计量,形成特征集向量,通过特征集的距离比较,在小尺寸水印质量退化情况下,实现了对水印标识的识别.分析和实验表明,该方案能够满足无线多媒体消息版权管理的需求,并且在文中分析的情况下,设备的自动识别精度可以达到95%以上,较好地支持了对无线多媒体消息的版权管理.

安全策略模型聚合性评估方法

动态策略支持与授权粒度是访问控制的关键问题.现有的研究只关注安全策略的描述能力,却忽略了对策略结构与授权粒度的分析,从而无法全面满足动态策略支持与最小授权要求.指出Lampson访问矩阵模型是对最细粒度访问控制的抽象,普通安全策略则根据应用安全需求对Lampson访问矩阵进行聚合.基于安全标签的聚合性描述框架(a descriptive framework of groupability basing on security labels,简称GroSeLa)可将普通安全策略映射为Lampson访问矩阵,该框架分为基本组件与扩展两部分:前者分析用于实现矩阵聚合的安全策略结构;后者则指出实现全面动态策略支持必须支持的7类管理性需求.在此基础上,提出5项聚合性指标:聚合因子、动态因子、策略规模、授权粒度与职责隔离支持.对4类经典安全策略ACL,BLP,DTE与RBAC的评估,是从矩阵聚合的角度分析不同的安全策略在表达性、可用性与授权粒度上的差异.

Decision tree classification of land cover from remotely sensed data

Decision tree classification algorithms have significant potential for land cover mapping problems and have not been tested in detail by the remote sensing community relative to more conventional pattern recognition techniques such as maximum likelihood classification. In this paper, we present several types of decision tree classification algorithms arid evaluate them on three different remote sensing data sets. The decision tree classification algorithms tested include an univariate decision tree, a multivariate decision tree, and a hybrid decision tree capable of including several different types of classification algorithms within a single decision tree structure. Classification accuracies produced by each of these decision tree algorithms are compared with both maximum likelihood and linear discriminant function classifiers. Results from this analysis show that the decision tree algorithms consistently outperform the maximum likelihood and linear discriminant function classifiers in regard to classf — cation accuracy. In particular, the hybrid tree consistently produced the highest classification accuracies for the data sets tested. More generally, the results from this work show that decision trees have several advantages for remote sensing applications by virtue of their relatively simple, explicit, and intuitive classification structure. Further, decision tree algorithms are strictly nonparametric and, therefore, make no assumptions regarding the distribution of input data, and are flexible and robust with respect to nonlinear and noisy relations among input features and class labels.

Semisupervised Dimensionality Reduction and Classification Through Virtual Label Regression

Semisupervised dimensionality reduction has been attracting much attention as it not only utilizes both labeled and unlabeled data simultaneously, but also works well in the situation of out-of-sample. This paper proposes an effective approach of semisupervised dimensionality reduction through label propagation and label regression. Different from previous efforts, the new approach propagates the label information from labeled to unlabeled data with a well-designed mechanism of random walks, in which outliers are effectively detected and the obtained virtual labels of unlabeled data can be well encoded in a weighted regression model. These virtual labels are thereafter regressed with a linear model to calculate the projection matrix for dimensionality reduction. By this means, when the manifold or the clustering assumption of data is satisfied, the labels of labeled data can be correctly propagated to the unlabeled data; and thus, the proposed approach utilizes the labeled and the unlabeled data more effectively than previous work. Experimental results are carried out upon several databases, and the advantage of the new approach is well demonstrated.

A electrochemiluminescence aptasensor for detection of thrombin incorporating the capture aptamer labeled with gold nanoparticles immobilized onto the thio-silanized ITO electrode

A novel electrochemiluminescence (ECL) aptasensor was proposed for sensitive and cost-effective detection of the target thrombin adopted an aptamer-based sandwich format. To detect thrombin, capture aptamers; labeled with gold nanoparticles (AuNPs) were first immobilized onto the thio-silanized ITO electrode surface through strong Au-S bonds. After catching the target thrombin, signal aptamers; tagged with ECL labels were attached to the assembled electrode surface. As a result, an AuNPs-capture-aptamer/thrombin/ECL-tagged signal-aptamer sandwich type was formed.

A study of resonance electron capture ionization on a quadrupole tandem mass spectrometer

Procedures that allow the realization of resonance electron capture (REC) mode on a commercial triple-quadrupole mass spectrometer, after some simple modifications, are described, REC mass spectrometry (MS) and tandem mass spectrometry (MS/MS) experiments were performed and spectra for some compounds were recorded. In particular, the charge-remote fragmentation (CRF) spectra of [M - H](-) ions of docosanoic and docosenoic acids under low-energy collisionally activated dissociation (CAD) conditions were obtained, and showed that there were no significant differences for [M - H](-) ions produced at different resonances (i,e. for [M - H](-) ions with different structures). This observation was explained on the basis of results obtained from deuterium-labeled fatty acids, which showed that different CRF ions (but with the same m/z value in the absence of labels) could be produced by different mechanisms, and all of them were obviously realized under CAD conditions that made spectra practically indistinguishable. The other example, which compared the REC-MS/MS spectrum of [M - H](-) ions and EI-MS/MS spectrum of M+. ions of daidzein, demonstrated the potential of the REC-MS/MS technique for more complex structure elucidation. Copyright (C) 2000 John Wiley & Sons, Ltd.

3-5岁幼儿认知灵活性的发展

To test preschoolers’ development of cognitive flexibility--an ability to solve a problem in one way and to then switch solution strategies, and the mechanism involved in the development, 3-5-year-olds are asked to perform switching tasks in which the experimenter manipulates the way the stimuli are presented: consecutive or simultaneous; the way the switching happens: between dimensions or within a dimension; the conceptual domains involved: shape, color, number and direction; the specific labels used. The main results of this work are presented below: (1) 3-5-year-olds’ cognitive flexibility develops with age, yet its development is not of the same speed in extra-dimensional switch tasks and inter-dimensional reversal tasks. 3-year-olds manifest some cognitive flexibility, but their performance is significantly worse than that of 4- and 5-year-olds. For the 3-year-olds, in reversal tasks, although 80% of the children passed the post-switch phrase in color task; less then 60% children passed the post-switch phrase in shape, number and direction tasks. In extra-dimensional tasks, 3-year-olds performance is worse than that in the reversal tasks. Less than 50% of the children passed the tasks. Children’s cognitive flexibility develops fast from 3-year-olds to 4-year-olds. Both 4-year-olds and 5-year-olds demonstrate high flexibility without significant difference between them. (2) Children’s flexibility in the conceptual domains of shape, color, number and direction follows different developing patterns. In inter-dimensional reversal tasks, 3-year-olds’ performance is not the same in the 4 conceptual domains, but the difference among the domains is insignificant in 4-and-5-year-olds. In extra-dimensional switching tasks, children’s performance on the 4 domain tasks is significantly different from one another in 3-, 4-, and 5-year-olds. (3) The way the stimuli are presented affects children’s development of cognitive flexibility. In inter-dimensional reversal tasks, 3-year-olds’ performance in consecutive presentation is significantly better than that in simultaneous presentation. 4- and 5-year-olds’ performance in the 2 presentations is not significantly different from each other. In extra-dimensional switch tasks, 3-, 4-, and 5-year-olds’ performance in the consecutive presentation is not significantly better than that in the simultaneous presentation (4) 3-, 4-, and 5-year-olds’ self-issued labeling aids their performance on the switching tasks. Children’ performance in the labeling condition is significantly better than that of no labeling. (5) 3-5-year-olds’ cognitive flexibility is highly correlated with their working memory and inhibition. Children’ development of cognitive flexibility is a process that involves activation of working memory and inhibition, in which the complexity of the task also plays a role.

Permutation Tests for Classification

We introduce and explore an approach to estimating statistical significance of classification accuracy, which is particularly useful in scientific applications of machine learning where high dimensionality of the data and the small number of training examples render most standard convergence bounds too loose to yield a meaningful guarantee of the generalization ability of the classifier. Instead, we estimate statistical significance of the observed classification accuracy, or the likelihood of observing such accuracy by chance due to spurious correlations of the high-dimensional data patterns with the class labels in the given training set. We adopt permutation testing, a non-parametric technique previously developed in classical statistics for hypothesis testing in the generative setting (i.e., comparing two probability distributions). We demonstrate the method on real examples from neuroimaging studies and DNA microarray analysis and suggest a theoretical analysis of the procedure that relates the asymptotic behavior of the test to the existing convergence bounds.

How smokers may react to cigarette taxes and price increases in Brazil: data from a national survey

Background: Despite being the third largest tobacco producer in the world, Brazil has developed a comprehensive tobacco control policy that includes a broad restriction on both advertising and smoking in indoor public places, compulsory pictorial warning labels, and a menthol cigarette ban. However, tax and pricing policies have been developed slowly and only very recently were stronger measures implemented. This study investigated the expected responses of smokers to hypothetical price increases in Brazil.Methods: We analyzed smokers' responses to hypothetical future price increases according to sociodemographic characteristics and smoking conditions in a multistage sample of Brazilian current cigarette smokers aged >= 14 years (n = 500). Logistic regression analysis was used to examine the relationship between possible responses and different predictors.Results: in most subgroups investigated, smokers most frequently said they would react to a hypothetical price increase by taking up alternatives that might have a positive impact on health, i.e., they would try to stop smoking (52.3%) or smoke fewer cigarettes (46.8%). However, a considerable percentage responded that they would use alternatives that would reduce the effect of price increases, such as the same brand with lower cost (48.1%). After controlling for sex age group (14-19, 20-39, 40-59, and >= 60 years), schooling level (>= 9 versus <= 9 years), number of cigarettes per day (>20 versus <= 20), and stage of change for smoking cessation (precontemplation, contemplation, and preparation), lower levels of dependence were positively associated with the response I would try to stop smoking (odds ratio [OR], 2.19). Young age was associated with I would decrease the number of cigarettes (OR, 3.44). A low schooling level was strongly associated with all responses.Conclusions: Taxes and prices increases have great potential to stimulate cessation or reduction of cigarette consumption further among two important vulnerable populations of smokers in Brazil: young smokers and those of low educational level. the results from the present study also suggest that seeking illegal products may reduce the impact of increased taxes, but does not eliminate it.