Bayes factor based speaker clustering for speaker diarization

This paper proposes the use of the Bayes Factor to replace the Bayesian Information Criterion (BIC) as a criterion for speaker clustering within a speaker diarization system. The BIC is one of the most popular decision criteria used in speaker diarization systems today. However, it will be shown in this paper that the BIC is only an approximation to the Bayes factor of marginal likelihoods of the data given each hypothesis. This paper uses the Bayes factor directly as a decision criterion for speaker clustering, thus removing the error introduced by the BIC approximation. Results obtained on the 2002 Rich Transcription (RT-02) Evaluation dataset show an improved clustering performance, leading to a 14.7% relative improvement in the overall Diarization Error Rate (DER) compared to the baseline system.

Bayes Factor based speaker segmentation for speaker diarization

This paper proposes the use of the Bayes Factor as a distance metric for speaker segmentation within a speaker diarization system. The proposed approach uses a pair of constant sized, sliding windows to compute the value of the Bayes Factor between the adjacent windows over the entire audio. Results obtained on the 2002 Rich Transcription Evaluation dataset show an improved segmentation performance compared to previous approaches reported in literature using the Generalized Likelihood Ratio. When applied in a speaker diarization system, this approach results in a 5.1% relative improvement in the overall Diarization Error Rate compared to the baseline.

Modelos para dados grupados e sensurados aplicados à área biológica: comparação usando fator de Bayes

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Automatic speaker recognition under adverse conditions

Speaker verification is the process of verifying the identity of a person by analysing their speech. There are several important applications for automatic speaker verification (ASV) technology including suspect identification, tracking terrorists and detecting a person’s presence at a remote location in the surveillance domain, as well as person authentication for phone banking and credit card transactions in the private sector. Telephones and telephony networks provide a natural medium for these applications. The aim of this work is to improve the usefulness of ASV technology for practical applications in the presence of adverse conditions. In a telephony environment, background noise, handset mismatch, channel distortions, room acoustics and restrictions on the available testing and training data are common sources of errors for ASV systems. Two research themes were pursued to overcome these adverse conditions: Modelling mismatch and modelling uncertainty. To directly address the performance degradation incurred through mismatched conditions it was proposed to directly model this mismatch. Feature mapping was evaluated for combating handset mismatch and was extended through the use of a blind clustering algorithm to remove the need for accurate handset labels for the training data. Mismatch modelling was then generalised by explicitly modelling the session conditions as a constrained offset of the speaker model means. This session variability modelling approach enabled the modelling of arbitrary sources of mismatch, including handset type, and halved the error rates in many cases. Methods to model the uncertainty in speaker model estimates and verification scores were developed to address the difficulties of limited training and testing data. The Bayes factor was introduced to account for the uncertainty of the speaker model estimates in testing by applying Bayesian theory to the verification criterion, with improved performance in matched conditions. Modelling the uncertainty in the verification score itself met with significant success. Estimating a confidence interval for the "true" verification score enabled an order of magnitude reduction in the average quantity of speech required to make a confident verification decision based on a threshold. The confidence measures developed in this work may also have significant applications for forensic speaker verification tasks.

Speaker diarization : "who spoke when"

Speaker diarization is the process of annotating an input audio with information that attributes temporal regions of the audio signal to their respective sources, which may include both speech and non-speech events. For speech regions, the diarization system also specifies the locations of speaker boundaries and assign relative speaker labels to each homogeneous segment of speech. In short, speaker diarization systems effectively answer the question of ‘who spoke when’. There are several important applications for speaker diarization technology, such as facilitating speaker indexing systems to allow users to directly access the relevant segments of interest within a given audio, and assisting with other downstream processes such as summarizing and parsing. When combined with automatic speech recognition (ASR) systems, the metadata extracted from a speaker diarization system can provide complementary information for ASR transcripts including the location of speaker turns and relative speaker segment labels, making the transcripts more readable. Speaker diarization output can also be used to localize the instances of specific speakers to pool data for model adaptation, which in turn boosts transcription accuracies. Speaker diarization therefore plays an important role as a preliminary step in automatic transcription of audio data. The aim of this work is to improve the usefulness and practicality of speaker diarization technology, through the reduction of diarization error rates. In particular, this research is focused on the segmentation and clustering stages within a diarization system. Although particular emphasis is placed on the broadcast news audio domain and systems developed throughout this work are also trained and tested on broadcast news data, the techniques proposed in this dissertation are also applicable to other domains including telephone conversations and meetings audio. Three main research themes were pursued: heuristic rules for speaker segmentation, modelling uncertainty in speaker model estimates, and modelling uncertainty in eigenvoice speaker modelling. The use of heuristic approaches for the speaker segmentation task was first investigated, with emphasis placed on minimizing missed boundary detections. A set of heuristic rules was proposed, to govern the detection and heuristic selection of candidate speaker segment boundaries. A second pass, using the same heuristic algorithm with a smaller window, was also proposed with the aim of improving detection of boundaries around short speaker segments. Compared to single threshold based methods, the proposed heuristic approach was shown to provide improved segmentation performance, leading to a reduction in the overall diarization error rate. Methods to model the uncertainty in speaker model estimates were developed, to address the difficulties associated with making segmentation and clustering decisions with limited data in the speaker segments. The Bayes factor, derived specifically for multivariate Gaussian speaker modelling, was introduced to account for the uncertainty of the speaker model estimates. The use of the Bayes factor also enabled the incorporation of prior information regarding the audio to aid segmentation and clustering decisions. The idea of modelling uncertainty in speaker model estimates was also extended to the eigenvoice speaker modelling framework for the speaker clustering task. Building on the application of Bayesian approaches to the speaker diarization problem, the proposed approach takes into account the uncertainty associated with the explicit estimation of the speaker factors. The proposed decision criteria, based on Bayesian theory, was shown to generally outperform their non- Bayesian counterparts.

Recursive pathways to marginal likelihood estimation with prior-sensitivity analysis

We investigate the utility to computational Bayesian analyses of a particular family of recursive marginal likelihood estimators characterized by the (equivalent) algorithms known as "biased sampling" or "reverse logistic regression" in the statistics literature and "the density of states" in physics. Through a pair of numerical examples (including mixture modeling of the well-known galaxy dataset) we highlight the remarkable diversity of sampling schemes amenable to such recursive normalization, as well as the notable efficiency of the resulting pseudo-mixture distributions for gauging prior-sensitivity in the Bayesian model selection context. Our key theoretical contributions are to introduce a novel heuristic ("thermodynamic integration via importance sampling") for qualifying the role of the bridging sequence in this procedure, and to reveal various connections between these recursive estimators and the nested sampling technique.

Bayesian predictive modeling and comparison of oil samples

Statistical comparison of oil samples is an integral part of oil spill identification, which deals with the process of linking an oil spill with its source of origin. In current practice, a frequentist hypothesis test is often used to evaluate evidence in support of a match between a spill and a source sample. As frequentist tests are only able to evaluate evidence against a hypothesis but not in support of it, we argue that this leads to unsound statistical reasoning. Moreover, currently only verbal conclusions on a very coarse scale can be made about the match between two samples, whereas a finer quantitative assessment would often be preferred. To address these issues, we propose a Bayesian predictive approach for evaluating the similarity between the chemical compositions of two oil samples. We derive the underlying statistical model from some basic assumptions on modeling assays in analytical chemistry, and to further facilitate and improve numerical evaluations, we develop analytical expressions for the key elements of Bayesian inference for this model. The approach is illustrated with both simulated and real data and is shown to have appealing properties in comparison with both standard frequentist and Bayesian approaches

Computational methods for Bayesian model choice

In this note, we shortly survey some recent approaches on the approximation of the Bayes factor used in Bayesian hypothesis testing and in Bayesian model choice. In particular, we reassess importance sampling, harmonic mean sampling, and nested sampling from a unified perspective.

Local disorder investigation in NiS2-xSex using Raman and Ni K-edge x-ray absorption spectroscopies

We report on Raman and Ni K-edge x-ray absorption investigations of a NiS2-xSex (with x = 0.00, 0.50/0.55, 0.60, and 1.20) pyrite family. The Ni K-edge absorption edge shows a systematic shift going from an insulating phase (x = 0.00 and 0.50) to a metallic phase (x = 0.60 and 1.20). The near-edge absorption features show a clear evolution with Se doping. The extended x-ray absorption fine structure data reveal the evolution of the local structure with Se doping which mainly governs the local disorder. We also describe the decomposition of the NiS2-xSex Raman spectra and investigate the weights of various phonon modes using Gaussian and Lorentzian profiles. The effectiveness of the fitting models in describing the data is evaluated by means of Bayes factor estimation. The Raman analysis clearly demonstrates the disorder effects due to Se alloying in describing the phonon spectra of NiS2-xSex pyrites.

Value of information in product recovery decisions: A Bayesian approach

There is a widespread recognition of the need for better information sharing and provision to improve the viability of end-of-life (EOL) product recovery operations. The emergence of automated data capture and sharing technologies such as RFID, sensors and networked databases has enhanced the ability to make product information; available to recoverers, which will help them make better decisions regarding the choice of recovery option for EOL products. However, these technologies come with a cost attached to it, and hence the question 'what is its value?' is critical. This paper presents a probabilistic approach to model product recovery decisions and extends the concept of Bayes' factor for quantifying the impact of product information on the effectiveness of these decisions. Further, we provide a quantitative examination of the factors that influence the value of product information, this value depends on three factors: (i) penalties for Type I and Type II errors of judgement regarding product quality; (ii) prevalent uncertainty regarding product quality and (iii) the strength of the information to support/contradict the belief. Furthermore, we show that information is not valuable under all circumstances and derive conditions for achieving a positive value of information. © 2010 Taylor & Francis.

Cosmological constrains on fast transition Unified Dark Matter models

In this work we show the results obtained applying a Unified Dark Matter (UDM) model with a fast transition to a set of cosmological data. Two different functions to model the transition are tested, and the feasibility of both models is explored using CMB shift data from Planck [1], Galaxy Clustering data from [2] and [3], and Union2.1 SNe Ia [4]. These new models are also statistically compared with the ACDM and quiessence models using Bayes factor through evidence. Bayesian inference does not discard the UDM models in favor of ACDM.

Molecular evidence for diphyly of the Asian catfish family Amblycipitidae (Teleostei: Siluriformes) and exclusion of the South American Aspredinidae from Sisoroidea

Amblycipitidae Day, 1873 is an Asian family of catfishes (Siluriformes) usually considered to contain 28 species placed in three genera: Amblyceps (14 spp.), Liobagrus (12 spp.) and Xiurenbagrus (2 spp.). Morphology-based systematics has supported the monophyly of this family, with some authors placing Amblycipitidae within a larger group including Akysidae, Sisoridae and Aspredinidae, termed the Sisoroidea. Here we investigate the phylogenetic relationships among four species of Amblyceps, six species of Liobagrus and the two species of Xiurenbagrus with respect to other sisoroid taxa as well as other catfish groups using 6100 aligned base pairs of DNA sequence data from the rag1 and rag2 genes of the nuclear genome and from three regions (cyt b, COL ND4 plus tRNA-His and tRNA-Ser) of the mitochondrial genome. Parsimony and Bayesian analyses of the data indicate strong support for a diphyletic Amblycipitidae in which the genus Amblyceps is the sister group to the Sisoridae and a clade formed by genera Liobagrus and Xiurenbagrus is the sister group to Akysidae. These taxa together form a well supported monophyletic group that assembles all Asian sisoroid taxa, but excludes the South American Aspredinidae. Results for aspredinids are consistent with previous molecular studies that indicate these catfishes are not sisoroids, but the sister group to the South American doradoid catfishes (Auchenipteridae + Doradidae). The redefined sisoroid clade plus Bagridae, Horabagridae and (Ailia + Laides) make up a larger monophyletic group informally termed "Big Asia." Likelihood-based SH tests and Bayes Factor comparisons of the rag and the mitochondrial data partitions considered separately and combined reject both the hypothesis of amblycipitid monophyly and the hypothesis of aspredinid inclusion within Sisoroidea. This result for amblycipitids conflicts with a number of well documented morphological synapomorphies that we briefly review. Possible nomenclatural changes for amblycipitid taxa are noted.