Browsing Faculty of Engineering by Author "Barnard, Etienne"
Now showing items 1-20 of 28
-
Adapting mobile medical information search to low-resourced areas
(IST-Africa, 2013)Providing good medical care in low-resourced areas is a challenge faced by many low and middle income countries. Continuously improving mobile communication infrastructure in these areas is however providing the opportunity ... -
Benign interpolation of noise in deep learning
(South African Institute of Computer Scientists and Information Technologists, 2020)The understanding of generalisation in machine learning is in a state of flux, in part due to the ability of deep learning models to interpolate noisy training data and still perform appropriately on out-of-sample data, ... -
Classifying recognised speech with deep neural networks
(Southern African Conference for Artificial Intelligence Research, 2020)We investigate whether word embeddings using deep neural networks can assist in the analysis of text produced by a speechrecognition system. In particular, we develop algorithms to identify which words are incorrectly ... -
Collecting and evaluating speech recognition corpora for 11 South African languages
(Springer, 2011)We describe the Lwazi corpus for automatic speech recognition (ASR), a new telephone speech corpus which contains data from the eleven official languages of South Africa. Because of practical constraints, the amount of ... -
Correlation between rapid learnability and user preference in IVR systems for developing regions
(iIST-Africa, 2013)Access to information and communication is one of the most important needs in any population group. It is generally challenging for people in the developing world to access information because the tools and the technologies ... -
Cross-bandwidth adaptation for ASR systems
(Pattern recognition association of South Africa (PRASA), 2013)Mismatches between application and training data greatly reduce the performance of automatic speech recognition (ASR) systems. However, collecting suitable amounts of in-domain and application-specific data for training ... -
Determination and the no-free-lunch paradox
(MIT Press, 2011)We discuss the no-free-lunch NFL theorem for supervised learning as a logical paradox—that is, as a counterintuitive result that is correctly proven from apparently incontestable assumptions. We show that the uniform prior ... -
DNNs as layers of cooperating classifiers
(AAAI Press, 2020-01)A robust theoretical framework that can describe and predict the generalization ability of deep neural networks (DNNs) in general circumstances remains elusive. Classical attempts have produced complexity metrics that rely ... -
Effects of application type on the choice of interaction modality in IVR systems
(Unisa Press, 2012)This paper addresses the feasibility of using the telephone as a tool for information access in the technology challenged and illiterate communities of Southern Africa. We did two case studies of disparate Interactive Voice ... -
Efficient data selection for ASR
(Language Resources and Evaluation, 2015)Automatic speech recognition (ASR) technology has matured over the past few decades and has made significant impacts in a variety of fields, from assistive technologies to commercial products. However, ASR system development ... -
Exploring minimal pronunciation modeling for low resource languages
(IOS Press Inc, 2015)Pronunciation lexicons can range from fully graphemic (modeling each word using the orthography directly) to fully phonemic (first mapping each word to a phoneme string). Between these two options lies a continuum of ... -
Generating fundamental frequency contours for speech synthesis in Yorùbá
(International Speech Communication Association ( ISCA ), 2013)We present methods for modelling and synthesising fundamental frequency (F0) contours suitable for application in text-to-speech (TTS) synthesis of Yorùbá (an African tone language). These methods are discussed and compared ... -
Improved transition models for cepstral trajectories
(Pattern recognition association of South Africa (PRASA), 2012)We improve on a piece-wise linear model of the trajectories of Mel Frequency Cepstral Coefficients, which are commonly used as features in Automatic Speech Recognition. For this purpose, we have created a very clean ... -
Insights regarding overfitting on noise in deep learning
(In Proc. South African Forum for Artificial Intelligence Research (FAIR2019), 2019-12)The understanding of generalization in machine learning is in a state of flux. This is partly due to the relatively recent revelation that deep learning models are able to completely memorize training data and still perform ... -
Kernel bandwidth estimation for non-parametric density estimation: a comparative study
(Pattern recognition association of South Africa (PRASA), 2013)We investigate the performance of conventional bandwidth estimators for non- parametric kernel density estimation on a number of representative pattern-recognition tasks, to gain a better understanding of the behaviour of ... -
Medium-vocabulary speech recognition for under-resourced languages
(SLTU, 2012)We report on the development of speech-recognition systems that are able to perform accurate recognition on mediumvocabulary tasks (i.e. tasks that require distinctions between approximately 200 different terms). We are ... -
Minimum phase finite impulse response filter design
(Wiley, 2022)The design of minimum phase finite impulse response (FIR) filters is considered. The study demonstrates that the residual errors achieved by current state‐of‐the‐art design methods are nowhere near the smallest error ... -
Optimising word embeddings for recognised multilingual speech
(Southern African Conference for Artificial Intelligence Research, 2020)Word embeddings are widely used in natural language processing (NLP) tasks. Most work on word embeddings focuses on monolingual languages with large available datasets. For embeddings to be useful in a multilingual ... -
Predicting utterance pitch targets in Yoruba for tone realisation in speech synthesis
(Elsevier, 2014)Pitch is a fundamental acoustic feature of speech and as such needs to be determined during the process of speech synthesis. While a range of communicative functions are attributed to pitch variation in speech of all ... -
ReLU and sigmoidal activation functions
(In Proc. South African Forum for Artificial Intelligence Research (FAIR2019), 2019-12)The generalization capabilities of deep neural networks are not well understood, and in particular, the influence of activation functions on generalization has received little theoretical attention. Phenomena such as ...
