- 138 West Street, Sandown, Sandton, Gauteng, South Africa 2191
- Call Us +27 76 238 2733
Menu
Development of the softer elements of linguistics-based software
Polyglot addresses the need to access content amongst the many possible formats available to store data digitally. Polyglot also addresses the problem of information loss that inevitably occurs through conversion and provides means of quantifying that loss and then minimizing it during future conversions.
They saw a gap in the market to provide services that research language and dialects in South Africa and thereby use their rigorous linguistic techniques to evaluate existing linguistics toolset for South African languages and improve on them. This in turn has helped them to develop language and dialect mapping for South Africa using state-of-the-art GIS software.
Our more people-centred approach involving volunteers much in the way that Wikipedia has been created is likely to yield more robust results than other more automated approaches.
Polyglot follows the philosophy that open-source software is more robust than proprietary software because it faces greater scrutiny from a larger group of human stakeholders.
Collaboration makes data management and repository building happen faster and more efficiently, when does right. And when this happens across social media there is no better source.
Polyglot’s strategy is to map all South Africa’s eleven official languages (L11) and non-official languages (non-L11). This will in turn help them towards developing an African Language Solutions Programme. The company also plans to develop a unique and patentable data-acquisition process that enables the embedding of location for both L11 and non-L11 audio files, thus enabling the deployment of sound collection resources and L11 and non-L11 audio files.
Critical to the success of a large portion of Polyglot programmes is the collection of sound data relevant to selected African Language Solutions Programme languages. To do this thorough planning will be developed on the collection process and locations of audio files and the deployment of sound collection resources to collect L11 and non-L11 languages. The collected audio data and sound-to-text processing will then be evaluated and refined.
This project will enable the team to perform detailed segmental phonology of pronunciation as well as detailed super-segmental phonology of pronunciation, analysis of combined elements of phonology of pronunciation and finally to compile final audio dataset and analysis documentation.
This programme will be formed from a compilation of Complex and Ambiguous Named Entity Recognition for both languages. And finally, all information will be combined to form tagging repositories for the L11 and non-L11 languages.
Another key output of Polyglot’s research will involve the creation of an Ontology of Knowledge algorithm for L11 and non-L11 languages, which will then enable it to develop linguistic sensitivity ‘key phrases’ and user-sensitivity adjustments toolset based on socio-demographic and geographical user data.