GENERAL ENQUIRIES: Tel: + 27 12 841 2911 | Email: callcentre@csir.co.za

Browsing by Author "Barnard, E"

Sort by: Order: Results:

  • Van Heerden, C; Barnard, E; Davel, M (International Speech Communication Association, 2009-09)
    Spoken dialogue systems (SDSs) have great potential for information access in the developing world. However, the realisation of that potential requires the solution of several challenging problems, including the development ...
  • Van Dyk, E; Barnard, E (PRASA 2008, 2008-11)
    We investigate the use of Naive Bayesian classifiers for correlated Gaussian feature spaces and derive error estimates for these classifiers. The error analysis is done by developing an exact expression for the error ...
  • Davel, MH; Barnard, E (PRASA, 2003-11)
    This paper describes a method for improving the efficiency of the language resource development process through bootstrapping.
  • Davel, MH; Barnard, E (Interspecch 2005, 2005-09)
    Bootstrapping techniques are an efficient way to develop electronic pronunciation dictionaries, but require fast system response to be practical for medium-to-large lexicons. In addition, user errors are inevitable during ...
  • Davel, M; Barnard, E (Acad Science South Africa A S S AF, 2006-07)
    Bootstrapping techniques can accelerate the development of language technology for resource-scarce languages. We (the authors) define a framework for the analysis of a general bootstrapping process whereby a model is ...
  • Davel, M; Barnard, E (2006-02-27)
    Bootstrapping techniques can accelerate the development of language technology for new languages. The authors define a framework for the analysis of a general bootstrapping process whereby a model is improved through a ...
  • Barnard, E (PRASA, 2009-11)
    The authors have previously argued that the infamous "No Free Lunch" theorem for supervised learning is a paradoxical result of a misleading choice of prior probabilities. Here, they provide more analysis of the dangers ...
  • Kleynhans, N; Barnard, E (PRASA 2008, 2008-11)
    The performance of trainable speech-processing systems deteriorates significantly when there is a mismatch between the training and testing data. The data mismatch becomes a dominant factor when collecting speech data for ...
  • Badenhorst, J; Van Heerden, C; Davel, M; Barnard, E (Springer Science+Business Media B.V., 2011-08)
    The authors 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 ...
  • Badenhorst, JAC; Van Heerden, C; Davel, M; Barnard, E (Association for Computational Linguistics, 2009-03)
    The authors describes the Lwazi corpus for automatic speech recognition (ASR), a new telephone speech corpus which includes data from nine Southern Bantu languages. Because of practical constraints, the amount of speech ...
  • Van Heerden, C; Barnard, E (PRASA 2009, 2009-11)
    The authors compare several different classifier combination methods on a single task, namely speaker age classification. This task is well suited to combination strategies, since significantly different feature classes ...
  • Loots, L; Davel, M; Barnard, E; Niesler, T (PRASA 2009, 2009-11)
    Phoneme-to-phoneme (P2P) learning provides a mechanism for predicting the pronunciation of a word based on its pronunciation in a different accent, dialect or language. The authors evaluate the effectiveness of manually-developed ...
  • Sharma Grover, A; Barnard, E (Bibliotheca Alexandrina, 2011-05)
    Over the past decade applications of speech technologies for development (ST4D) have shown much potential for enabling information access and service delivery. In this paper the authors review two deployed ST4D services ...
  • Govender, N; Kuun, C; Zimu, V; Barnard, E; Davel, M (2006-04)
    Two related issues were investigated in the computational modeling of Nguni prosody, based on annotated databases of isiZulu and isiXhosa speech. Firstly, authors show that a simple template can be used to describe the ...
  • Smit, WJ; Barnard, E (Elsevier, 2009-04)
    Sparse coding is an efficient way of coding information. In a sparse code most of the code elements are zero; very few are active. Sparse codes are intended to correspond to the spike trains with which biological neurons ...
  • Ndwe, TJ; Barnard, E; Foko, Thato E (IST Africa 2013 Conference Proceedings, 2013-05)
    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 ...
  • Kleynhans, N; Barnard, E (PRASA 2013 Proceedings, 2013-12)
    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 ...
  • Van der Walt, Christiaan M; Barnard, E (2006-11)
    The relationship between the distribution of data, on the one hand, and classifier performance, on the other, for non-parametric classifiers has been studied. It is shown that predictable factors such as the available ...
  • Davel, MH; Barnard, E (PRASA 2004, 2004-11)
    The authors define a novel g-to-p prediction algorithm that utilises the concept of a 'default phoneme': a grapheme which is realised as a specific phoneme significantly more often than as any other phoneme. They found ...
  • Van der Walt, Christiaan M; Barnard, E (PRASA, 2009-11)
    The authors propose a hyper-ellipsoid clustering algorithm that grows clusters from local structures in a dataset and estimates the underlying geometrical structure of data with a set of hyper-ellipsoids. The clusters are ...