The days of men dominating the engineering scene is still an evident and alarming fact. Despite an improvement in recent years, there remains a serious shortage of women pursuing careers in engineering. The North-West University (NWU) aims to change this - one semester at a time…
The NWU’s Faculty of Engineering will once again host the annual Femmegineering celebration on Saturday, 3 August 2019 where the attending grade 10, 11 and 12 girls can look forward to speakers who specialise in various interesting and somewhat complicated fields.
SADiLaR, the South African Centre for Digital Language Resources (funded by the Department of Science and Technology and hosted by NWU), will also be present at this celebration, and is excited to have two guest speakers at the event.
Tanja Gaustad, the manager for resource development and senior computational linguist at CTexT, will give an introduction to human language technology based on the example of word prediction in language modelling and its real-world use in applications. The focus will be on the diversity of tasks and techniques, along with the reason why it could be interesting for someone with a more technical interest in engineering, programming and/or intelligent systems.
“Computational linguistics is an interdisciplinary field of research where natural language (spoken by humans) is modelled from a computational perspective, allowing for the analysis of large amounts of data,” she says.
“What makes this field interesting and rewarding to me is, on the one hand, working in interdisciplinary teams with linguists, programmers and native speakers of a language, and on the other hand, that there are still so many unsolved problems for less common languages to explore and improve.
Dr Febe de Wet, yet another expert in the field of digital signal processing at Stellenbosch University, will also address the audience about speech-to-text-systems.
“The people of South Africa are multilingual: the country recognises 11 official languages, and in practice even more are spoken,” she says. “Automatic speech recognisers (speech-to-text systems) are however designed for monolingual input, as may be expected in many large economies such as North America, Japan or countries in Europe. In contrast, everyday South African speech often contains a mixture of languages and such mixed language speech cannot be understood by even the best commercial speech recognisers available today.
“To develop a mixed-language speech recognition system, a data set of mixed-language speech is needed,” she says.
Dr de Wet is currently working on a project in which a data set containing such speech in five South African languages is being developed.
If you have not applied yet, we invite you to apply to attend the annual Femmegineering celebration on the NWU’s campus in Potchefstroom.
Spaces are limited, so selection for attendance is based on grades for maths and science along with a motivation. Applying does not guarantee attendance. Successful applicants will receive an e-mail or SMS.
Contact Annika Pretorius at email@example.com for more details or visit this link to register: http://iqi.su/cVsVb