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

Human language technologies

 

News

CSIR Research Space is constantly being developed and populated. Please report any errors to Siphethile Muswelanto. Constructive criticism and suggestions would be welcome.

Recent Submissions

  • Schlünz, Georg I; Wilken, Ilana; Moors, Carmen; Gumede, Tebogo; Van der Walt, Willem J; Calteaux, Karen V; Tönsing, K; Van Niekerk, K (ACM Digital Library, 2017-09)
    Persons with certain disabilities face barriers to information access and interpersonal communication. Assistive technologies provide workaround solutions to these problems. Augmentative and alternative communication systems ...
  • Marais, Laurette; Van Zijl, L (Springer Nature, 2016-07)
    We investigate self-verifying nondeterministic finite automata, in the case of unary symmetric difference nondeterministic finite automata (SV-XNFA). We show that there is a family of languages Ln=2 which can always be ...
  • Govender, Avashna; De Wet, Febe; Tapamo, Jules-Raymond (Technische Universität Berlin, 2015-09)
    Hidden Markov Model (HMM)-based synthesis in combination with speaker adaptation has proven to be an approach that is well-suited for child speech synthesis. This paper describes the development and evaluation of different ...
  • Govender, N; Nouhou, B; De Wet, F (IEEE, 2015-11)
    Acquiring large amounts of child speech data is a particularly difficult task. One could therefore consider the possibility to add existing corpora of child speech data to the severely limited resources that are available ...
  • Marais, L; Pretorius, L (Southern African Computer Lecturers Association (SACLA), 2015-07)
    Higher education in South Africa is, amongst others, characterised by student language diversity while the language of teaching and learning is mainly English. The University of South Africa has a language policy that is ...
  • Badenhorst, Jaco; Tshoane, Alfred; De Wet, Febe (IEEE, 2016-12)
    This paper reports on an analysis of isiXhosa speech produced by adult language learners. The learners whose speech was recorded were all acquiring isiXhosa as an additional language and the majority of the students had ...
  • Govender, Avashna; De Wet, Febe (IEEE, 2016-12)
    Building synthetic child voices is considered a difficult task due to the challenges associated with data collection. As a result, speaker adaptation in conjunction with Hidden Markov Model (HMM)-based synthesis has become ...
  • De Wet, Febe; Badenhorst, J; Modipa, T (Elsevier, 2016-05)
    The official languages of South Africa can still be classified as under-resourced with respect to the speech resources that are required for technology development. Harvesting speech data from existing sources is one means ...
  • Badenhorst, J; Davel, MH (IEEE, 2015-11)
    We experiment with a new method to create synthetic models of rare and unseen triphones in order to supplement limited automatic speech recognition (ASR) training data. A trajectory model is used to characterise seen ...
  • Wilken, I; De Wet, Febe; Taljard, E (ISCA, 2015-09)
    Communication plays a vital role in everyday life and in some situations multilingual communication is a necessity. The South African constitution recognises 11 official languages. Multilingual communication therefore ...
  • Sahraeian, R; Van Compernolle, D; De Wet, Febe (IEEE, 2015-11)
    Recently, multilingual deep neural networks (DNNs) have been successfully used to improve under-resourced speech recognizers. Common approaches use either a merged universal phoneme set based on the International Phonetic ...
  • De Wet, Febe; Marais, L; Klop, D (ISCA Archive, 2015-09)
    The purpose of this study was to measure the efficacy of an eBook to improve the vocabulary and word recognition skills in an Afrikaans speaking group of lower socio-economic status of 6- to 7-year old children with poor ...
  • Heeringa, W; De Wet, Febe; Van Huyssteen, GB (University of Stellenbosch. Department of General Linguistics, 2015-07)
    Following Den Besten’s (2009) desiderata for historical linguistics of Afrikaans, this article aims to contribute some modern evidence to the debate regarding the founding dialects of Afrikaans. From an applied perspective ...
  • Kleynhans, N; De Wet, Febe; Barnard, E (IEEE, 2015-11)
    Large amounts of untranscribed audio data are generated every day. These audio resources can be used to develop robust acoustic models that can be used in a variety of speech-based systems. Manually transcribing this data ...
  • Strik, H; Palumbo, L; De Wet, Febe; Cucchiarini, C (International Speech Communication Association, 2015-09)
    The European ‘Lifelong Learning Programme’ (LLP) project ‘Games Online for Basic Language learning’ (GOBL) aimed to provide youths and adults wishing to improve their basic language skills access to materials for the ...
  • Sahraeian, R; van Compernolle, D; De Wet, Febe (International Speech Communication Association, 2015-09)
    Conventional acoustic modeling involves estimating many parameters to effectively model feature distributions. The sparseness of speech and text data, however, degrades the reliability of the estimation process and makes ...
  • Manqele, L; Dlodlo, M; Coetzee, L; Williams, Q; Sibiya, G (IEEE, 2015-09)
    The usage of the Internet of Things technology across different service provisioning environments has increased the challenges associated with service discovery and selection. Users cannot always remember the Internet ...
  • Burke, Michael G; Sabatta, D (PRASA, 2015-11)
    Conference scheduling and organising is a particularly laborious task and can be exteremely time consuming. While many online conference platforms allow manual topic selection, these can be expensive and typically still ...
  • Kleynhans, NT; Barnard, E (Springer Verlag, 2014-10)
    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 ...
  • Kleynhans, N; De Wet, Febe (Pattern Recognition Association of South Africa, 2014-11)
    Most of the developing world can still be classified as under-resourced in terms of their languages. Harvesting suitable and relatively easily accessible spoken resources can drastically improve the situation. One such ...

View more