Improving automatic speaker verification using SVM techniques

Automatic recognition of people is an active field of research with important forensic and security applications. In these applications, it is not always possible for the subject to be in close proximity to the system. Voice represents a human behavioural trait which can be used to recognise people in such situations. Automatic Speaker Verification (ASV) is the process of verifying a persons identity through the analysis of their speech and enables recognition of a subject at a distance over a telephone channel { wired or wireless. A significant amount of research has focussed on the application of Gaussian mixture model (GMM) techniques to speaker verification systems providing state-of-the-art performance. GMM's are a type of generative classifier trained to model the probability distribution of the features used to represent a speaker. Recently introduced to the field of ASV research is the support vector machine (SVM). An SVM is a discriminative classifier requiring examples from both positive and negative classes to train a speaker model. The SVM is based on margin maximisation whereby a hyperplane attempts to separate classes in a high dimensional space. SVMs applied to the task of speaker verification have shown high potential, particularly when used to complement current GMM-based techniques in hybrid systems. This work aims to improve the performance of ASV systems using novel and innovative SVM-based techniques. Research was divided into three main themes: session variability compensation for SVMs; unsupervised model adaptation; and impostor dataset selection. The first theme investigated the differences between the GMM and SVM domains for the modelling of session variability | an aspect crucial for robust speaker verification. Techniques developed to improve the robustness of GMMbased classification were shown to bring about similar benefits to discriminative SVM classification through their integration in the hybrid GMM mean supervector SVM classifier. Further, the domains for the modelling of session variation were contrasted to find a number of common factors, however, the SVM-domain consistently provided marginally better session variation compensation. Minimal complementary information was found between the techniques due to the similarities in how they achieved their objectives. The second theme saw the proposal of a novel model for the purpose of session variation compensation in ASV systems. Continuous progressive model adaptation attempts to improve speaker models by retraining them after exploiting all encountered test utterances during normal use of the system. The introduction of the weight-based factor analysis model provided significant performance improvements of over 60% in an unsupervised scenario. SVM-based classification was then integrated into the progressive system providing further benefits in performance over the GMM counterpart. Analysis demonstrated that SVMs also hold several beneficial characteristics to the task of unsupervised model adaptation prompting further research in the area. In pursuing the final theme, an innovative background dataset selection technique was developed. This technique selects the most appropriate subset of examples from a large and diverse set of candidate impostor observations for use as the SVM background by exploiting the SVM training process. This selection was performed on a per-observation basis so as to overcome the shortcoming of the traditional heuristic-based approach to dataset selection. Results demonstrate the approach to provide performance improvements over both the use of the complete candidate dataset and the best heuristically-selected dataset whilst being only a fraction of the size. The refined dataset was also shown to generalise well to unseen corpora and be highly applicable to the selection of impostor cohorts required in alternate techniques for speaker verification.

Mechanisms of change in negative thinking and urinary monoamines in depressed patients during acute treatment with group cognitive behavior therapy and antidepressant medication

This naturalistic study investigated the mechanisms of change in measures of negative thinking and in 24-h urinary metabolites of noradrenaline (norepinephrine), dopamine and serotonin in a sample of 43 depressed hospital patients attending an eight-session group cognitive behavior therapy program. Most participants (91%) were taking antidepressant medication throughout the therapy period according to their treating Psychiatrists' prescriptions. The sample was divided into outcome categories (19 Responders and 24 Non-responders) on the basis of a clinically reliable change index [Jacobson, N.S., & Truax, P., 1991. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19.] applied to the Beck Depression Inventory scores at the end of the therapy. Results of repeated measures analysis of variance [ANOVA] analyses of variance indicated that all measures of negative thinking improved significantly during therapy, and significantly more so in the Responders as expected. The treatment had a significant impact on urinary adrenaline and metadrenaline excretion however, these changes occurred in both Responders and Non-responders. Acute treatment did not significantly influence the six other monoamine metabolites. In summary, changes in urinary monoamine levels during combined treatment for depression were not associated with self-reported changes in mood symptoms.

Negative mood regulation expectancies, frontal lobe related behaviors and alcohol use

Negative mood regulation (NMR) expectancies have been linked to substance problems in previous research, but the neurobiological correlates of NMR are unknown. In the present study, NMR was examined in relation to self-report indices of frontal lobe functioning, mood and alcohol use in 166 volunteers of both genders who ranged in age from 17 to 43 years. Contrary to expectations based on previous findings in addicts and problem drinkers, scores on the NMR scale did not differ between Low Risk and High Risk drinkers as defined by the Alcohol Use Disorders Identification Test (AUDIT). However, NMR scores were significantly negatively correlated with all three indices of frontal lobe dysfunction on the Frontal Systems Behavior Scale (FrSBe) Self-Rating Form as well as with all three indices of negative mood on the Depression Anxiety Stress Scales (DASS), which in turn were all positively correlated with FrSBe. Path analyses indicated that NMR partially mediated the direct effects of frontal lobe dysfunction (as indexed by FrSBe) on DASS Stress and DASS Depression. Further, the High Risk drinkers scored significantly higher on the Disinhibition and Executive Dysfunction indices of the FrSBe than did Low Risk drinkers. Results are consistent with the notion that NMR is a frontal lobe function.

The role of social capital in reducing negative health outcomes among police officers

This paper investigates the role of social capital on the reduction of short and long run negative health effects associated with stress, as well as indicators of burnout among police officers. Despite the large volume of research on either social capital or the health effects of stress, the interaction of these factors remains an underexplored topic. In this empirical analysis we aim to reduce such a shortcoming focusing on a highly stressful and emotionally draining work environment, namely law enforcement agents who perform as an essential part of maintaining modern society. Using a multivariate regression analysis focusing on three different proxies of health and three proxies for social capital conducting also several robustness checks, we find strong evidence that increased levels of social capital is highly correlated with better health outcomes. Additionally we observe that while social capital at work is very important, social capital in the home environment and work-life balance are even more important. From a policy perspective, our findings suggest that work and stress programs should actively encourage employees to build stronger social networks as well as incorporate better working/home life arrangements.

Distance Education and the Integration of E-learning in a Graduate Program

Historically, distance education consisted of a combination of face-to-face blocks of time and surface mailed packages. However, advances in information technology literacy and the abundance of personal computers has placed e-learning in increased demand. The authors describe the planning, implementation, and evaluation of the blending of e-learning with face-to-face education in the postgraduate nursing forum. Experiences of this particular student group are also discussed.

Integration of a road traffic noise model (ASJ) and traffic simulation (AVENUE) for a built-up area

A road traffic noise prediction model (ASJ MODEL-1998) has been integrated with a road traffic simulator (AVENUE) to produce the Dynamic areawide Road traffic NoisE simulator-DRONE. This traffic-noise-GIS based integrated tool is upgraded to predict noise levels in built-up areas. The integration of traffic simulation with a noise model provides dynamic access to traffic flow characteristics and hence automated and detailed predictions of traffic noise. The prediction is not only on the spatial scale but also on temporal scale. The linkage with GIS gives a visual representation to noise pollution in the form of dynamic areawide traffic noise contour maps. The application of DRONE on a real world built-up area is also presented.

Forecasting crashes at the planning level: Simultaneous negative binomial crash model applied in Tucson, Arizona

At least two important transportation planning activities rely on planning-level crash prediction models. One is motivated by the Transportation Equity Act for the 21st Century, which requires departments of transportation and metropolitan planning organizations to consider safety explicitly in the transportation planning process. The second could arise from a need for state agencies to establish incentive programs to reduce injuries and save lives. Both applications require a forecast of safety for a future period. Planning-level crash prediction models for the Tucson, Arizona, metropolitan region are presented to demonstrate the feasibility of such models. Data were separated into fatal, injury, and property-damage crashes. To accommodate overdispersion in the data, negative binomial regression models were applied. To accommodate the simultaneity of fatality and injury crash outcomes, simultaneous estimation of the models was conducted. All models produce crash forecasts at the traffic analysis zone level. Statistically significant (p-values < 0.05) and theoretically meaningful variables for the fatal crash model included population density, persons 17 years old or younger as a percentage of the total population, and intersection density. Significant variables for the injury and property-damage crash models were population density, number of employees, intersections density, percentage of miles of principal arterial, percentage of miles of minor arterials, and percentage of miles of urban collectors. Among several conclusions it is suggested that planning-level safety models are feasible and may play a role in future planning activities. However, caution must be exercised with such models.

Sustainable urban and transport development for transportation disadvantaged : a review

Around the world, particularly in North America and Australia, urban sprawl combined with low density suburban development has caused serious accessibility and mobility problems, especially for those who do not own a motor vehicle or have access to public transportation services. Sustainable urban and transportation development is seen crucial in solving transportation disadvantage problems in urban settlements. However, current urban and transportation models have not been adequately addressed unsustainable urban transportation problems that transportation disadvantaged groups overwhelmingly encounter, and the negative impacts on the disadvantaged have not been effectively considered. Transportation disadvantaged is a multi-dimensional problem that combines demographic, spatial and transportation service dimensions. Nevertheless, most transportation models focusing on transportation disadvantage only employ demographic and transportation service dimensions and do not take spatial dimension into account. This paper aims to investigate the link between sustainable urban and transportation development and spatial dimension of the transportation disadvantage problem. The paper, for that purpose, provides a thorough review of the literature and identifies a set of urban, development and policy characteristics to define spatial dimension of the transportation disadvantage problem. This paper presents an overview of these urban, development and policy characteristics that have significant relationships with sustainable urban and transportation development and travel inability, which are also useful in determining transportation disadvantaged populations.

Forecasting negative effects of monotony and sensation seeking on performance during a vigilance task

The driving task requires sustained attention during prolonged periods, and can be performed in highly predictable or repetitive environments. Such conditions could create hypovigilance and impair performance towards critical events. Identifying such impairment in monotonous conditions has been a major subject of research, but no research to date has attempted to predict it in real-time. This pilot study aims to show that performance decrements due to monotonous tasks can be predicted through mathematical modelling taking into account sensation seeking levels. A short vigilance task sensitive to short periods of lapses of vigilance called Sustained Attention to Response Task is used to assess participants‟ performance. The framework for prediction developed on this task could be extended to a monotonous driving task. A Hidden Markov Model (HMM) is proposed to predict participants‟ lapses in alertness. Driver‟s vigilance evolution is modelled as a hidden state and is correlated to a surrogate measure: the participant‟s reactions time. This experiment shows that the monotony of the task can lead to an important decline in performance in less than five minutes. This impairment can be predicted four minutes in advance with an 86% accuracy using HMMs. This experiment showed that mathematical models such as HMM can efficiently predict hypovigilance through surrogate measures. The presented model could result in the development of an in-vehicle device that detects driver hypovigilance in advance and warn the driver accordingly, thus offering the potential to enhance road safety and prevent road crashes.