Towards improved speech recognition for resource poor languages

In recent times, the improved levels of accuracy obtained by Automatic Speech Recognition (ASR) technology has made it viable for use in a number of commercial products. Unfortunately, these types of applications are limited to only a few of the world’s languages, primarily because ASR development is reliant on the availability of large amounts of language specific resources. This motivates the need for techniques which reduce this language-specific, resource dependency. Ideally, these approaches should generalise across languages, thereby providing scope for rapid creation of ASR capabilities for resource poor languages. Cross Lingual ASR emerges as a means for addressing this need. Underpinning this approach is the observation that sound production is largely influenced by the physiological construction of the vocal tract, and accordingly, is human, and not language specific. As a result, a common inventory of sounds exists across languages; a property which is exploitable, as sounds from a resource poor, target language can be recognised using models trained on resource rich, source languages. One of the initial impediments to the commercial uptake of ASR technology was its fragility in more challenging environments, such as conversational telephone speech. Subsequent improvements in these environments has gained consumer confidence. Pragmatically, if cross lingual techniques are to considered a viable alternative when resources are limited, they need to perform under the same types of conditions. Accordingly, this thesis evaluates cross lingual techniques using two speech environments; clean read speech and conversational telephone speech. Languages used in evaluations are German, Mandarin, Japanese and Spanish. Results highlight that previously proposed approaches provide respectable results for simpler environments such as read speech, but degrade significantly when in the more taxing conversational environment. Two separate approaches for addressing this degradation are proposed. The first is based on deriving better target language lexical representation, in terms of the source language model set. The second, and ultimately more successful approach, focuses on improving the classification accuracy of context-dependent (CD) models, by catering for the adverse influence of languages specific phonotactic properties. Whilst the primary research goal in this thesis is directed towards improving cross lingual techniques, the catalyst for investigating its use was based on expressed interest from several organisations for an Indonesian ASR capability. In Indonesia alone, there are over 200 million speakers of some Malay variant, provides further impetus and commercial justification for speech related research on this language. Unfortunately, at the beginning of the candidature, limited research had been conducted on the Indonesian language in the field of speech science, and virtually no resources existed. This thesis details the investigative and development work dedicated towards obtaining an ASR system with a 10000 word recognition vocabulary for the Indonesian language.

An integrated rail track degradation model

There has been a worldwide trend to increase axle loads and train speeds. This means that railway track degradation will be accelerated, and track maintenance costs will be increased significantly. There is a need to investigate the consequences of increasing traffic load. The aim of the research is to develop a model for the analysis of physical degradation of railway tracks in response to changes in traffic parameters, especially increased axle loads and train speeds. This research has developed an integrated track degradation model (ITDM) by integrating several models into a comprehensive framework. Mechanistic relationships for track degradation hav~ ?een used wherever possible in each of the models contained in ITDM. This overcc:mes the deficiency of the traditional statistical track models which rely heavily on historical degradation data, which is generally not available in many railway systems. In addition statistical models lack the flexibility of incorporating future changes in traffic patterns or maintenance practices. The research starts with reviewing railway track related studies both in Australia and overseas to develop a comprehensive understanding of track performance under various traffic conditions. Existing railway related models are then examined for their suitability for track degradation analysis for Australian situations. The ITDM model is subsequently developed by modifying suitable existing models, and developing new models where necessary. The ITDM model contains four interrelated submodels for rails, sleepers, ballast and subgrade, and track modulus. The rail submodel is for rail wear analysis and is developed from a theoretical concept. The sleeper submodel is for timber sleepers damage prediction. The submodel is developed by modifying and extending an existing model developed elsewhere. The submodel has also incorporated an analysis for the likelihood of concrete sleeper cracking. The ballast and subgrade submodel is evolved from a concept developed in the USA. Substantial modifications and improvements have been made. The track modulus submodel is developed from a conceptual method. Corrections for more global track conditions have been made. The integration of these submodels into one comprehensive package has enabled the interaction between individual track components to be taken into account. This is done by calculating wheel load distribution with time and updating track conditions periodically in the process of track degradation simulation. A Windows-based computer program ~ssociated with ITDM has also been developed. The program enables the user to carry out analysis of degradation of individual track components and to investigate the inter relationships between these track components and their deterioration. The successful implementation of this research has provided essential information for prediction of increased maintenance as a consequence of railway trackdegradation. The model, having been presented at various conferences and seminars, has attracted wide interest. It is anticipated that the model will be put into practical use among Australian railways, enabling track maintenance planning to be optimized and potentially saving Australian railway systems millions of dollars in operating costs.

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.