This research study deals mainly with the influence of the medium of instruction on throughput at the Vaal University of Technology. The underlying hypothesis driving this research, is that learners at the Vaal University of Technology will tend to underachieve during their period of study, largely because the medium of instruction and assessment is not their first language, but a second or even a third language. The assertion is that learners on tertiary level can only perform academically well if they have gained a certain level of competency in their mother tongue, to ease the difficult process of acquiring the second language English, which is the main medium of instruction at the VUT. For most human beings language is the medium through which knowledge is transferred or negotiated. The success of this interaction is determined by the effectiveness of communication. Thus it would be fair to say, that if the means of communication is inappropriate, there will be little or no language transfer. Further, it seems logical that before any other considerations are made with regard to teaching and learning, the instrument which enables this interaction should first be in place. The research aims are: to determine the influence of a second language as medium of instruction on academic performance, to establish the nature of the support system which might facilitate the advancement of learners not prepared for the academic demands of university, to determine what can be done to address the problem and provide possible recommendations for improved academic performance. In order to attain the abovementioned research aims, a literature review and an empirical investigation were undertaken. The literature study discussed the role which the medium of instruction has played in the history of South African Education. The South African Language Policy, before and after 1994, was then reviewed. Parent and learner choice in respect of the choice of medium of instruction was then considered. The findings of the empirical study have shown that the medium of instruction has a remarkable influence on learners' proficiency and eventual performance in their content subjects. The empirical research was conducted by using a questionnaire in order to obtain data on relevant variables as indicated by the literature study, as well as identify barriers perceived by students that hamper their academic progress. Descriptive statistics, such as frequencies, were used to summarize the data. Marks from the compulsory language test at the institution were also taken into account in order to establish learners' English language proficiency. This data, together with students' marks, were interpreted in the empirical analysis. The target population for this study consisted of L1 and L2 students who registered in 2004 and will have completed their studies in 2006, as well as students having completed in 2007. The population consisted of both female and male students from the Faculty of Visual Arts and Design. In the last chapter, Chapter 6, conclusions from the literature review and empirical investigation were drawn. Recommendations for further research were provided which stressed the need for developing an appropriate training course for L2MI (Second Language Medium of Instruction) content subject teachers. Effective training in L2MI is one of the most important factors in improving the level of academic literacy in South African learners. / Thesis (M.Ed.)--North-West University, Vaal Triangle Campus, 2009.

Both human perceptual studies and automaticmachine-based experiments have shown that visual information from a speaker's mouth region can improve the robustness of automatic speech processing tasks, especially in the presence of acoustic noise. By taking advantage of the complementary nature of the acoustic and visual speech information, audio-visual speech processing (AVSP) applications can work reliably in more real-world situations than would be possible with traditional acoustic speech processing applications. The two most prominent applications of AVSP for viable human-computer-interfaces involve the recognition of the speech events themselves, and the recognition of speaker's identities based upon their speech. However, while these two fields of speech and speaker recognition are closely related, there has been little systematic comparison of the two tasks under similar conditions in the existing literature. Accordingly, the primary focus of this thesis is to compare the suitability of general AVSP techniques for speech or speaker recognition, with a particular focus on synchronous hidden Markov models (SHMMs). The cascading appearance-based approach to visual speech feature extraction has been shown to work well in removing irrelevant static information from the lip region to greatly improve visual speech recognition performance. This thesis demonstrates that these dynamic visual speech features also provide for an improvement in speaker recognition, showing that speakers can be visually recognised by how they speak, in addition to their appearance alone. This thesis investigates a number of novel techniques for training and decoding of SHMMs that improve the audio-visual speech modelling ability of the SHMM approach over the existing state-of-the-art joint-training technique. Novel experiments are conducted within to demonstrate that the reliability of the two streams during training is of little importance to the final performance of the SHMM. Additionally, two novel techniques of normalising the acoustic and visual state classifiers within the SHMM structure are demonstrated for AVSP. Fused hidden Markov model (FHMM) adaptation is introduced as a novel method of adapting SHMMs from existing wellperforming acoustic hidden Markovmodels (HMMs). This technique is demonstrated to provide improved audio-visualmodelling over the jointly-trained SHMMapproach at all levels of acoustic noise for the recognition of audio-visual speech events. However, the close coupling of the SHMM approach will be shown to be less useful for speaker recognition, where a late integration approach is demonstrated to be superior.

Σκοπός της παρούσας διατριβής είναι η εκμετάλλευση των κυματιδίων με σκοπό την βελτίωση της απόδοσης συστημάτων αναγνώρισης ομιλητή και ομιλίας. Στα πλαίσια αυτά, εισάγονται τέσσερις νέοι τρόποι παραμετροποίησης του σήματος ομιλίας: (1) Η πρώτη μέθοδος προσαρμόζει την ανάλυση συχνότητας των πακέτων κυματιδίων για την προσέγγιση της ψυχοακουστικής επίδρασης των κρίσιμων ζωνών του ακουστικού συστήματος ενσωματώνοντας τις τελευταίες εξελίξεις για τον υπολογισμό τους. (2) Η δεύτερη μέθοδος εισάγει μια επέκταση του μετασχηματισμού πακέτων κυματιδίων, τον επικαλυπτόμενο μετασχηματισμό πακέτων κυματιδίων, ο οποίος χρησιμοποιείται για να δοθεί έμφαση στις περιοχές αλλαγής των κρίσιμων ζωνών από μια μικρότερη σε μια μεγαλύτερη τιμή. (3) Η τρίτη μέθοδος αξιολογεί τη συνεισφορά μη επικαλυπτόμενων ζωνών συχνοτήτων στην αναγνώριση ομιλητή και κατασκευάζεται ανάλογα ένας μετασχηματισμός πακέτων κυματιδίων ο οποίος προσαρμόζει την συχνοτική του ανάλυση σύμφωνα με την απόδοση κάθε μίας από τις ζώνες. (4) Η τέταρτη μέθοδος επιλέγει τη βέλτιστη βάση από το σύνολο των μετασχηματισμών που είναι διαθέσιμοι με τα πακέτα κυματιδίων με εφαρμογή την αναγνώριση ομιλητή και κριτήριο το μέτρο EER. Οι παραπάνω τέσσερις τρόποι παραμετροποίησης του σήματος ομιλίας αξιολογήθηκαν με το σύστημα αναγνώρισης ομιλητή WCL-1 του εργαστηρίου ενσύρματης τηλεπικοινωνίας του Πανεπιστημίου Πατρών στις βάσεις δεδομένων POLYCOST και NIST και αποδείχθηκε η ανωτερότητά τους τόσο σε σχέση με προηγούμενες μεθόδους των κυματιδίων όσο και σε σχέση με ευρέως χρησιμοποιούμενες παραμέτρους ομιλίας, όπως οι παράμετροι cepstral με βάση την κλίμακα mel (MFCC). Επιπλέον, στη διατριβή αναλύονται οι ιδιότητες των σημαντικότερων συναρτήσεων κυματιδίων, επιλέγεται η βέλτιστη για την αναπαράσταση του σήματος ομιλίας και πιστοποιείται στην πράξη αυτή η επιλογή. Τέλος, οι δύο πρώτες από τις προαναφερόμενες μεθόδους παραμετροποίησης τροποποιήθηκαν και επεκτάθηκαν κατάλληλα για την εφαρμογή στην αναγνώριση ομιλίας όπου αξιολογήθηκαν και διαπιστώθηκε η υπεροχή τους έναντι παραδοσιακών και ευρέως διαδεδομένων μεθόδων παραμετροποίησης του σήματος ομιλίας που στηρίζονται στον μετασχηματισμό Fourier. Το κύριο συμπέρασμα που προέκυψε από τη παρούσα διδακτορική διατριβή είναι ότι τα κυματίδια και συγκεκριμένα τα πακέτα κυματιδίων είναι δυνατόν να χρησιμοποιηθούν με επιτυχία στη βελτίωση της απόδοσης συστημάτων αναγνώρισης ομιλητή και ομιλίας. / The main goal of the present thesis is the exploitation of wavelets for the optimization of speaker and speech recognition systems performance. In this context, four new speech parameterization methods are introduced: (1) The first method adapts the frequency resolution of wavelet packet transform to the critical bandwidth of auditory filters incorporating the recent advances for their estimation. (2) The second method introduces a generalization of wavelet packet transform, named overlapping wavelet packet transform, which emphasizes those frequency sub-bands that critical bandwidth changes from a finer to a coarser value. (3) The third method evaluates the contribution of each one of eight non-overlapping frequency sub-bands, that the Nyquist interval is divided, to the speaker recognition task and a wavelet packet transform is constructed which adapts its frequency resolution according to the performance of each sub-band. (4) The fourth method introduces a new technique for seeking and selecting the best basis among all wavelet packet transforms available in the speaker recognition task taking as criterion the EER. The aforementioned four speech signal parameterizations were evaluated on the speaker verification system WCL-1 of Wire Communications Laboratory, University of Patras, utilizing the speaker recognition corpora POLYCOST and NIST and their superiority was proven over previous wavelet-based parameterizations as well as the widely used Mel Frequency Cepstral Coefficients (MFCC). Among the four proposed methods, it was proven that the second parameterization technique exhibited the best performance. Furthermore, the most important wavelet properties are thoroughly analyzed, the optimal is selected for the representation of the speech signal and this choice is experimentally verified. Finally, the first two parameterization methods were further modified and extended appropriately for application on the speech recognition task where their superiority was proven over traditionally and widely used speech parameterization techniques based on Fourier transform. The main conclusion that resulted in the present doctoral thesis is that wavelets and specifically wavelet packet transforms can be used successfully for the tasks of speaker and speech recognition.

La thèse s’intéresse à la problématique de la répétition dans le discours. Le langage de la répétition relève de la doxa, mot que nous avons utilisé dans la thèse comme le terme générique qui englobe cette réalité complexe que Nathalie Sarraute dénonce dans son œuvre en s’attaquant au réalisme discursif. Il s’agit d’étudier dans un corpus littéraire le langage figé qui exprime des réalités figées. C’est un langage qui s’appuie sur un système de référence prototypique. Le référent est soit un objet, soit une propriété ou un processus isolable dont les réalistes considèrent qu’il existe en dehors de notre esprit. C’est le discours de la modélisation qui privilégie ce que Paul VALERY appelle dans Monsieur Teste « la machine » de langage. Ce sont les habitudes langagières qui consistent à inventer des codes d’écriture et de lecture servant de règle à toutes les communautés doxiques dans leur rapport avec le monde. Ce langage apparaît à travers l’utilisation des formes génériques et figées comme le stéréotype, le lieu commun, le cliché, le préjugé, l’idée reçue. La thèse essaie de mettre en exergue les stratégies de dénonciation d’une telle forme de discours dans le roman de Nathalie Sarraute. Elle passe en revue l’énonciation et la dénonciation des stéréotypes qui se divisent en stéréotypes de pensées et en stéréotypes de langue. / The thesis deals with the problem of repetition in speech. The language of repetition is the doxa, word that we used in the thesis as the generic term that encompasses this complex reality that Nathalie Sarraute denounces in his work by attacking the discursive realism. He is studying the set language that expresses frozen realities in a literary corpus. It is a language that relies on a prototypical reference system. The referent is either an object, either a property or a reportable process wich realists consider that there are outside our mind. It is the speech of modeling that privileges what Paul VALERY call in Mr tests the "machine language ". These are the language habits which consist in inventing of the codes of writing and reading rule for all doxa communities in their relation to the world. This language appears through the use of generic and frozen forms as the stereotype, the common place, the cliché, the prejudice, the received idea. The thesis tries to highlight strategies for the reporting of such a form of speech in the novel to Nathalie Sarraute. She will review the enunciation and the denunciation of the stereotypes that divide in stereotypes of thoughts and language stereotypes

Some researchers that work with metaphor theory claim that metaphors and figurative language are understood and processed just as easily as literal language. However, as this thesis will explore in detail, other research indicates that such is not always the case. That is, if the category of metaphor is further subcategorized into conventional and non-conventional metaphor, the scope will change because of the fact that it is possible to argue that non-conventionalized metaphors require a more conscious path of processing. In order to explain this alternative path, there are two primary approaches to language processing worth introducing: implicit and explicit. These approaches vary in required attention and speed of processing. With regards to conscious effort, these approaches are rather similar to the way in which we process conventionalized and non-conventionalized metaphors. Conventional metaphors are processed more quickly and easily than non-conventional ones. Hence, the claim that all metaphors are similarly processed may not always be true. Furthermore, an individual’s level of proficiency presumably correlates with speed in language processing. However, if non-conventional metaphor requires a more deliberate path of processing, this thesis assumes that the processing of this type of metaphor will be relatively unaffected by proficiency level, thus causing informants to process them in similar manners. In this thesis, 24 non-native speakers (NNS), categorized into intermediate proficient and advanced proficient, and seven native speakers (NS) were tested with an RT-test on subjective metaphor comprehension. Results were compared using mean response times and standard deviations, as well as looking at correlations and coefficient of variation. The results showed a distinct difference in processing speed with conventional metaphors being processed significantly faster. Moreover, the findings indicate that conventional metaphor processing speed seems to be predicted by proficiency, whilst non-conventional processing speed is not. The RT differences remained relatively consistent in both conventional and non-conventional metaphor processing, but when taking correlations, variance and coefficient of variation into consideration, the findings indicate that these other factors help level out the differences in non-conventional metaphor processing in more subtle ways than simply by RT’s.

Tato diplomová práce se zabývá problematikou vzdáleného rozpoznávání mluvčích. V případě dat zachycených odlehlým mikrofonem se přesnost standardního rozpoznávání značně snižuje, proto jsem navrhl dva přístupy pro zlepšení výsledků. Prvním z nich je použití mikrofonního pole (záměrně rozestavené sady mikrofonů), které je schopné nasměrovat virtuální "paprsek" na pozici řečníka. Dále jsem prováděl adaptaci komponent systému (PLDA skórování a extraktoru i-vektorů). S využitím simulace pokojových podmínek jsem syntetizoval trénovací a testovací data ze standardní datové sady NIST 2010. Ukázal jsem, že obě techniky a jejich kombinace vedou k výraznému zlepšení výsledků. Dále jsem se zabýval společným určením identity a pozice mluvčího. Zatímco výsledky ve venkovním simulovaném prostředí (bez ozvěn) jsou slibné, výsledky z interiéru (s ozvěnami) jsou smíšené a vyžadují další prozkoumání. Na závěr jsem mohl systémem vyhodnotit omezené množství reálných dat získaných přehráním a záznamem nahrávek ve skutečné místnosti. Zatímco výsledky pro mužské nahrávky odpovídají simulaci, výsledky pro ženské nahrávky nejsou přesvědčivé a vyžadují další analýzu.

La segmentation et le regroupement en locuteurs (SRL) impliquent la détection des locuteurs dans un flux audio et les intervalles pendant lesquels chaque locuteur est actif, c'est-à-dire la détermination de ‘qui parle quand’. La première partie des travaux présentés dans cette thèse exploite une approche de modélisation du locuteur utilisant des clés binaires (BKs) comme solution à la SRL. La modélisation BK est efficace et fonctionne sans données d'entraînement externes, car elle utilise uniquement des données de test. Les contributions présentées incluent l'extraction des BKs basée sur l'analyse spectrale multi-résolution, la détection explicite des changements de locuteurs utilisant les BKs, ainsi que les techniques de fusion SRL qui combinent les avantages des BKs et des solutions basées sur un apprentissage approfondi. La tâche de la SRL est étroitement liée à celle de la reconnaissance ou de la détection du locuteur, qui consiste à comparer deux segments de parole et à déterminer s'ils ont été prononcés par le même locuteur ou non. Même si de nombreuses applications pratiques nécessitent leur combinaison, les deux tâches sont traditionnellement exécutées indépendamment l'une de l'autre. La deuxième partie de cette thèse porte sur une application où les solutions de SRL et de reconnaissance des locuteurs sont réunies. La nouvelle tâche, appelée détection de locuteurs à faible latence, consiste à détecter rapidement les locuteurs connus dans des flux audio à locuteurs multiples. Il s'agit de repenser la SRL en ligne et la manière dont les sous-systèmes de SRL et de détection devraient être combinés au mieux. / Speaker diarization (SD) involves the detection of speakers within an audio stream and the intervals during which each speaker is active, i.e. the determination of ‘who spoken when’. The first part of the work presented in this thesis exploits an approach to speaker modelling involving binary keys (BKs) as a solution to SD. BK modelling is efficient and operates without external training data, as it operates using test data alone. The presented contributions include the extraction of BKs based on multi-resolution spectral analysis, the explicit detection of speaker changes using BKs, as well as SD fusion techniques that combine the benefits of both BK and deep learning based solutions. The SD task is closely linked to that of speaker recognition or detection, which involves the comparison of two speech segments and the determination of whether or not they were uttered by the same speaker. Even if many practical applications require their combination, the two tasks are traditionally tackled independently from each other. The second part of this thesis considers an application where SD and speaker recognition solutions are brought together. The new task, coined low latency speaker spotting (LLSS), involves the rapid detection of known speakers within multi-speaker audio streams. It involves the re-thinking of online diarization and the manner by which diarization and detection sub-systems should best be combined.

La tâche de segmentation et de regroupement en locuteurs (speaker diarization) consiste à identifier "qui parle quand" dans un flux audio sans connaissance a priori du nombre de locuteurs ou de leur temps de parole respectifs. Les systèmes de segmentation et de regroupement en locuteurs sont généralement construits en combinant quatre étapes principales. Premièrement, les régions ne contenant pas de parole telles que les silences, la musique et le bruit sont supprimées par la détection d'activité vocale (VAD). Ensuite, les régions de parole sont divisées en segments homogènes en locuteur par détection des changements de locuteurs, puis regroupées en fonction de l'identité du locuteur. Enfin, les frontières des tours de parole et leurs étiquettes sont affinées avec une étape de re-segmentation. Dans cette thèse, nous proposons d'aborder ces quatre étapes avec des approches fondées sur les réseaux de neurones. Nous formulons d’abord le problème de la segmentation initiale (détection de l’activité vocale et des changements entre locuteurs) et de la re-segmentation finale sous la forme d’un ensemble de problèmes d’étiquetage de séquence, puis nous les résolvons avec des réseaux neuronaux récurrents de type Bi-LSTM (Bidirectional Long Short-Term Memory). Au stade du regroupement des régions de parole, nous proposons d’utiliser l'algorithme de propagation d'affinité à partir de plongements neuronaux de ces tours de parole dans l'espace vectoriel des locuteurs. Des expériences sur un jeu de données télévisées montrent que le regroupement par propagation d'affinité est plus approprié que le regroupement hiérarchique agglomératif lorsqu'il est appliqué à des plongements neuronaux de locuteurs. La segmentation basée sur les réseaux récurrents et la propagation d'affinité sont également combinées et optimisées conjointement pour former une chaîne de regroupement en locuteurs. Comparé à un système dont les modules sont optimisés indépendamment, la nouvelle chaîne de traitements apporte une amélioration significative. De plus, nous proposons d’améliorer l'estimation de la matrice de similarité par des réseaux neuronaux récurrents, puis d’appliquer un partitionnement spectral à partir de cette matrice de similarité améliorée. Le système proposé atteint des performances à l'état de l'art sur la base de données de conversation téléphonique CALLHOME. Enfin, nous formulons le regroupement des tours de parole en mode séquentiel sous la forme d'une tâche supervisée d’étiquetage de séquence et abordons ce problème avec des réseaux récurrents empilés. Pour mieux comprendre le comportement du système, une analyse basée sur une architecture de codeur-décodeur est proposée. Sur des exemples synthétiques, nos systèmes apportent une amélioration significative par rapport aux méthodes de regroupement traditionnelles. / Speaker diarization is the task of determining "who speaks when" in an audio stream that usually contains an unknown amount of speech from an unknown number of speakers. Speaker diarization systems are usually built as the combination of four main stages. First, non-speech regions such as silence, music, and noise are removed by Voice Activity Detection (VAD). Next, speech regions are split into speaker-homogeneous segments by Speaker Change Detection (SCD), later grouped according to the identity of the speaker thanks to unsupervised clustering approaches. Finally, speech turn boundaries and labels are (optionally) refined with a re-segmentation stage. In this thesis, we propose to address these four stages with neural network approaches. We first formulate both the initial segmentation (voice activity detection and speaker change detection) and the final re-segmentation as a set of sequence labeling problems and then address them with Bidirectional Long Short-Term Memory (Bi-LSTM) networks. In the speech turn clustering stage, we propose to use affinity propagation on top of neural speaker embeddings. Experiments on a broadcast TV dataset show that affinity propagation clustering is more suitable than hierarchical agglomerative clustering when applied to neural speaker embeddings. The LSTM-based segmentation and affinity propagation clustering are also combined and jointly optimized to form a speaker diarization pipeline. Compared to the pipeline with independently optimized modules, the new pipeline brings a significant improvement. In addition, we propose to improve the similarity matrix by bidirectional LSTM and then apply spectral clustering on top of the improved similarity matrix. The proposed system achieves state-of-the-art performance in the CALLHOME telephone conversation dataset. Finally, we formulate sequential clustering as a supervised sequence labeling task and address it with stacked RNNs. To better understand its behavior, the analysis is based on a proposed encoder-decoder architecture. Our proposed systems bring a significant improvement compared with traditional clustering methods on toy examples.

The aim of this project is to provide useful data from published dictionaries, corpora, and instructional materials, as well as sample lessons, to promote the teaching of Unorthodox Oral Expressions (UOEs) to learners of English as a second/foreign language. In the first chapter, the author reviews relevant literature, explains what UOEs are, and discusses the importance of incorporating UOEs in EFL or ESL classrooms. In the second chapter, a linguistic categorization of UOEs is given. In the third chapter, the results are given of an examination of 10 different dictionaries. The purpose of this examination was to find which of 56 target UOEs are included in each dictionary and what kind of definitions are given for them. The results show that many common UOEs are not included in most, or any, dictionaries. For the UOEs that are included in most dictionaries, the definitions do not always agree, and factors such as intonation are not taken into account. Moreover, the explanations on how the UOEs can be used are not complete. In the fourth chapter, three English language corpora are examined to discover which of the target 56 UOEs are the most frequently used. The results show some differences in UOE frequency between the corpora that include both spoken and written English text and the spoken English corpora. In the fifth chapter, the teaching of UOEs in ESL textbooks is analyzed. The results show that most of these books do not teach UOEs explicitly. In chapter six, experimental instructional units are provided. Results of piloting these lessons at Brigham Young University's English Language Center are discussed. In the last chapter, the author suggests possible future research involving UOEs.

No description available.