A new approach of speaker clustering based on the stereophonic differential energy

In this paper, we present a new approach of speech clustering with regards of the speaker identity. It consists in grouping the homogeneous speech segments that are obtained at the end of the segmentation process, by using the spatial information provided by the stereophonic speech signals. The proposed method uses the differential energy of the two stereophonic signals collected by two cardioid microphones, in order to cluster all the speech segments that belong to the same speaker. The total number of clusters obtained at the end should be equal to the real number of speakers present in the meeting room and each cluster should contain the global intervention of only one speaker. The proposed system is suitable for debates or multi-conferences for which the speakers are located at fixed positions. Basically, our approach tries to make a speaker localization with regards to the position of the microphones, taken as a spatial reference. Based on this localization, the new proposed method can recognize the speaker identity of any speech segment during the meeting. So, the intervention of each speaker is automatically detected and assigned to him by estimating his relative position. In a purpose of comparison, two types of clustering methods have been implemented and experimented: the new approach, which we called Energy Differential based Spatial Clustering (EDSC) and a classical statistical approach called “Mono-Gaussian based Sequential Clustering” (MGSC). Experiments of speaker clustering are done on a stereophonic speech corpus called DB15, composed of 15 stereophonic scenarios of about 3.5 minutes each. Every scenario corresponds to a free discussion between two or three speakers seated at fixed positions in the meeting room. Results show the outstanding performances of the new approach in terms of precision and speed, especially for short speech segments, where most of clustering techniques present a strong failure.     

On modelling non-probabilistic uncertainty in the likelihood ratio approach to evidential reasoning

When the likelihood ratio approach is employed for evidential reasoning in law, it is often necessary to employ subjective probabilities, which are probabilities derived from the opinions and judgement of a human (expert). At least three concerns arise from the use of subjective probabilities in legal applications. Firstly, human beliefs concerning probabilities can be vague, ambiguous and inaccurate. Secondly, the impact of this vagueness, ambiguity and inaccuracy on the outcome of a probabilistic analysis is not necessarily fully understood. Thirdly, the provenance of subjective probabilities and the associated potential sources of vagueness, ambiguity and inaccuracy tend to be poorly understood, making it difficult for the outcome of probabilistic reasoning to be explained and validated, which is crucial in legal applications. The former two concerns have been addressed by a wide body of research in AI. The latter, however, has received little attention. This paper presents a novel approach to employ argumentation to reason about probability distributions in probabilistic models. It introduces a range of argumentation schemes and corresponding sets of critical questions for the construction and validation of argument models that define sets of probability distributions. By means of an extended example, the paper demonstrates how the approach, argumentation schemes and critical questions can be employed for the development of models and their validation in legal applications of the likelihood ratio approach to evidential reasoning.     

Hierarchical clustering of subpopulations with a dissimilarity based on the likelihood ratio statistic: application to clustering massive data sets

The problem of clustering subpopulations on the basis of samples is considered within a statistical framework: a distribution for the variables is assumed for each subpopulation and the dissimilarity between any two populations is defined as the likelihood ratio statistic which compares the hypothesis that the two subpopulations differ in the parameter of their distributions to the hypothesis that they do not. A general algorithm for the construction of a hierarchical classification is described which has the important property of not having inversions in the dendrogram. The essential elements of the algorithm are specified for the case of well-known distributions (normal, multinomial and Poisson) and an outline of the general parametric case is also discussed. Several applications are discussed, the main one being a novel approach to dealing with massive data in the context of a two-step approach. After clustering the data in a reasonable number of ‘bins’ by a fast algorithm such as k-Means, we apply a version of our algorithm to the resulting bins. Multivariate normality for the means calculated on each bin is assumed: this is justified by the central limit theorem and the assumption that each bin contains a large number of units, an assumption generally justified when dealing with truly massive data such as currently found in modern data analysis. However, no assumption is made about the data generating distribution.

A clustering based approach for skyline diversity

Skyline query processing has recently received a lot of attention in database and data-mining communities. To the best of our knowledge, the existing researches mainly focus on considering how to efficiently return the whole skyline set. However, when the cardinality and dimensionality of input objects increase, the number of skylines grows exponentially, and hence this “huge” skyline set is completely useless to users. On the other hand, in most real applications, the objects are usually clustered, and therefore many objects have similar attribute values. Motivated by the above facts, in this paper, we present a novel type of SkyCluster query to capture the skyline diversity and improve the usefulness of skyline result. The SkyCluster query integrates K-means clustering into skyline computation, and returns K “representative” and “diverse” skyline objects to users. To process such query, a straightforward approach is to simply integrate the existing techniques developed for skyline-only and clustering-only together. But this approach is costly since both skyline computation and K-means clustering are all CPU-sensitive. We propose an efficient evaluation approach which is based on the circinal index to seamlessly integrate subspace skyline computation, K-means clustering and representatives selection. Also, we present a novel optimization heuristic to further improve the query performance. Experimental study shows that our approach is both efficient and effective.     

A Bayesian approach to model-based clustering for binary panel probit models

Considering latent heterogeneity is of special importance in nonlinear models in order to gauge correctly the effect of explanatory variables on the dependent variable. A stratified model-based clustering approach is adapted for modeling latent heterogeneity in binary panel probit models. Within a Bayesian framework an estimation algorithm dealing with the inherent label switching problem is provided. Determination of the number of clusters is based on the marginal likelihood and a cross-validation approach. A simulation study is conducted to assess the ability of both approaches to determine on the correct number of clusters indicating high accuracy for the marginal likelihood criterion, with the cross-validation approach performing similarly well in most circumstances. Different concepts of marginal effects incorporating latent heterogeneity at different degrees arise within the considered model setup and are directly at hand within Bayesian estimation via MCMC methodology. An empirical illustration of the methodology developed indicates that consideration of latent heterogeneity via latent clusters provides the preferred model specification over a pooled and a random coefficient specification.     

基于说话人模型聚类的说话人识别

为了提高说话人识别系统的识别效率,提出一种基于说话人模型聚类的说话人识别方法,通过近似KL距离将相似的说话人模型聚类,为每类确定类中心和类代表,构成分级说话人识别模型。测试时先通过计算测试矢量与类中心或类代表之间的距离选择类,再通过计算测试矢量与选中类中的说话人模型之间对数似然度确定目标说话人,这样可以大大减少计算量。实验结果显示,在相同条件下,基于说话人模型聚类的说话人识别的识别速度要比传统的GMM的识别速度快4倍,但是识别正确率只降低了0.95%。因此,与传统GMM相比,基于说话人模型聚类的说话人识别能在保证识别正确率的同时大大提高识别速度。     

A feature selection-based speaker clustering method for paralinguistic tasks

In recent years, computational paralinguistics has emerged as a new topic within speech technology. It concerns extracting non-linguistic information from speech (such as emotions, the level of conflict, whether the speaker is drunk). It was shown recently that many methods applied here can be assisted by speaker clustering; for example, the features extracted from the utterances could be normalized speaker-wise instead of using a global method. In this paper, we propose a speaker clustering algorithm based on standard clustering approaches like K-means and feature selection. By applying this speaker clustering technique in two paralinguistic tasks, we were able to significantly improve the accuracy scores of several machine learning methods, and we also obtained an insight into what features could be efficiently used to separate the different speakers.     

A new speaker verification method with global speaker model and likelihood score normalization

In this paper a new text-independent speaker verification method GSMSV is proposed based on likelihood score normalization.In this novel method a global speaker model is established to represent the universal features of speech and normalize the likelihood score.Statistical analysis demonstrates that this normalization method can remove common factors of speech and bring the differences between speakers into prominence.As a result the equal error rate is decreased significantly,verification procedure is accelerated and system adaptability to speaking speed is improved.     

说话人聚类的初始类生成方法

目前说话人聚类时将说话人分割后的语音段作为初始类,直接对这些数量庞大语音段进行聚类的计算量非常大。为了降低说话人聚类时的计算量,提出一种面向说话人聚类的初始类生成方法。提取说话人分割后语音段的特征参数及特征参数的质心,结合层次聚类法和贝叶斯信息准则,对语音段进行具有宽松停止准则的“预聚类”,生成初始类。与直接对说话人分割后的语音段进行聚类的方法相比,该方法能在保持原有聚类性能的情况下,减少40.04%的计算时间;在允许聚类性能略有下降的情形下,减少60.03%以上的计算时间。     

基于聚类的字典压缩技术在GIS中的应用研究

本文主要探讨了基于字典的矢量地图压缩中字典的设计问题,提出了一种新颖的基于聚类方式的字典设计方法,它可以使字典更好的近似于某种特定的数据集。实验证明只要字典结构适合,这种基于聚类方式的字典数据压缩技术可获得更好的压缩效果。     

Distribution-based anomaly detection via generalized likelihood ratio test: A general Maximum Entropy approach

We address the problem of detecting “anomalies” in the network traffic produced by a large population of end-users following a distribution-based change detection approach. In the considered scenario, different traffic variables are monitored at different levels of temporal aggregation (timescales), resulting in a grid of variable/timescale nodes. For every node, a set of per-user traffic counters is maintained and then summarized into histograms for every time bin, obtaining a timeseries of empirical (discrete) distributions for every variable/timescale node. Within this framework, we tackle the problem of designing a formal Distribution-based Change Detector (DCD) able to identify statistically-significant deviations from the past behavior of each individual timeseries.     

Flocking based approach for data clustering

Data clustering is a process of extracting similar groups of the underlying data whose labels are hidden. This paper describes different approaches for solving data clustering problem. Particle swarm optimization (PSO) has been recently used to address clustering task. An overview of PSO-based clustering approaches is presented in this paper. These approaches mimic the behavior of biological swarms seeking food located in different places. Best locations for finding food are in dense areas and in regions far enough from others. PSO-based clustering approaches are evaluated using different data sets. Experimental results indicate that these approaches outperform K-means, K-harmonic means, and fuzzy c-means clustering algorithms.     

说话人识别中基于聚类特征的矢量量化技术

为解决采用矢量量化的方法进行说话人识别时出现的失真问题,根据汉语语音的发音特性,提出了将矢量量化与语音特征的聚类技术相结合的方法,在进行矢量量化码书训练之前,先对特征矢量进行聚类筛选。实验结果表明,当测试语音片段长度为4 s时,在保持95％左右识别率下,采用普通矢量量化方法需64码本数,而采用该文方法只需8码本数,降低了8倍。结果说明该方法不但在一定程度上解决了因训练样本不足而引起的失真问题,而且通过方法的改进,实现了采用较低码字数产生较好的识别结果,从而提高识别效率。     

Control chart based on likelihood ratio for monitoring linear profiles

A control chart based on the likelihood ratio is proposed for monitoring the linear profiles. The new chart which integrates the EWMA procedure can detect shifts in either the intercept or the slope or the standard deviation, or simultaneously by a single chart which is different from other control charts in literature for linear profiles. The results by Monte Carlo simulation show that our approach has good performance across a wide range of possible shifts. We show that the new method has competitive performance relative to other methods in literature in terms of ARL, and another feature of the new chart is that it can be easily designed. The application of our proposed method is illustrated by a real data example from an optical imaging system.     

A local approach for Bayesian fusion: Mathematical analysis and agent based conception

An agent based architecture that is modelled on a successfully operating process of the real world–criminal investigation–circumvents high computational costs caused by Bayesian fusion by realising a distributed local Bayesian fusion approach. The idea underlying local Bayesian fusion approaches is to perform Bayesian fusion at least not in detail on the whole space that is spanned by the Properties-of-Interest. Local Bayesian fusion is mainly based on coarsening and restriction techniques. Here, we focus on coarsening. We give an overview over the agent based conception and translate the proposed ideas in a formal mathematical framework.     

Comparison of clustering methods: A case study of text-independent speaker modeling

Clustering is needed in various applications such as biometric person authentication, speech coding and recognition, image compression and information retrieval. Hundreds of clustering methods have been proposed for the task in various fields but, surprisingly, there are few extensive studies actually comparing them. An important question is how much the choice of a clustering method matters for the final pattern recognition application. Our goal is to provide a thorough experimental comparison of clustering methods for text-independent speaker verification. We consider parametric Gaussian mixture model (GMM) and non-parametric vector quantization (VQ) model using the best known clustering algorithms including iterative (K-means, random swap, expectation-maximization), hierarchical (pairwise nearest neighbor, split, split-and-merge), evolutionary (genetic algorithm), neural (self-organizing map) and fuzzy (fuzzy C-means) approaches. We study recognition accuracy, processing time, clustering validity, and correlation of clustering quality and recognition accuracy. Experiments from these complementary observations indicate clustering is not a critical task in speaker recognition and the choice of the algorithm should be based on computational complexity and simplicity of the implementation. This is mainly because of three reasons: the data is not clustered, large models are used and only the best algorithms are considered. For low-order models, choice of the algorithm, however, can have a significant effect.     

A clustering based approach to perceptual image hashing

A perceptual image hash function maps an image to a short binary string based on an image's appearance to the human eye. Perceptual image hashing is useful in image databases, watermarking, and authentication. In this paper, we decouple image hashing into feature extraction (intermediate hash) followed by data clustering (final hash). For any perceptually significant feature extractor, we propose a polynomial-time heuristic clustering algorithm that automatically determines the final hash length needed to satisfy a specified distortion. We prove that the decision version of our clustering problem is NP complete. Based on the proposed algorithm, we develop two variations to facilitate perceptual robustness versus fragility tradeoffs. We validate the perceptual significance of our hash by testing under Stirmark attacks. Finally, we develop randomized clustering algorithms for the purposes of secure image hashing.     

A new approach of clustering based machine-learning algorithm

Machine-learning research is to study and apply the computer modeling of learning processes in their multiple manifestations, which facilitate the development of intelligent system. In this paper, we have introduced a clustering based machine-learning algorithm called clustering algorithm system (CAS). The CAS algorithm is tested to evaluate its performance and find fruitful results. We have been presented some heuristics to facilitate machine-learning authors to boost up their research works. The InfoBase of the Ministry of Civil Services is used to analyze the CAS algorithm. The CAS algorithm is compared with other machine-learning algorithms like UNIMEM, COBWEB, and CLASSIT, and was found to have some strong points over them. The proposed algorithm combined advantages of two different approaches to machine learning. The first approach is learning from Examples, CAS supports Single and Multiple Inheritance and Exceptions. CAS also avoids probability assumptions which are well understood in concept formation. The second approach is learning by Observation. CAS applies a set of operators that have proven to be effective in conceptual clustering. We have shown how CAS builds and searches through a clusters hierarchy to incorporate or characterize an object.