Data Science for Social Impact (DSFSI)

🎉 A full-circle, unforgettable night at the "Science Oscars"! 🏆 We are absolutely thrilled to share that the TextAugment team has won the inaugural NSTF-SADiLaR Research Software Award for Human Language Technologies (HLT) at the prestigious NSTF-South32 Awards! At the Data Science for Social Impact (DSFSI) lab, we have always believed that software is not just a byproduct of research—it is the vital engine that drives it. 🚀 Why TextAugment Matters Data scarcity is one of the biggest hurdles preventing low-resource and African languages from fully participating in the modern AI revolution. TextAugment is an open-source library designed to generate synthetic training data, helping to build robust AI models for under-resourced languages. To date, the library has over 284,000+ lifetime downloads, with 2k–3k new downloads every month, powering research globally from Swahili and Arabic to Uzbek. 💻 GitHub: https://lnkd.in/efHU37Rm Read the paper https://lnkd.in/eAvmtNga 🤝 A Beautiful Full-Circle Moment The absolute highlight of the night was sharing the nominee stage with Dr. Herkulaas Combrink and his team at the University of the Free State (UFS), who were nominated for their vital work on South African Sign Language (SASL). As a former lab member, who Prof Vukosi Marivate had the privilege of co-supervising, seeing Herkulaas stand on that stage as a peer was incredibly special. It is living proof of our core philosophy: Ideas must travel, and they must travel with people. 💐 It Takes a Village Special thanks to: * Joseph Sefara (CSIR) and Isheanesu Dzingirai (UP) for keeping this engine running. * Council for Scientific and Industrial Research (CSIR) for the initial ignition spark, and African Institute for Mathematical Sciences (AIMS) South Africa for the pivotal research software workshop where this began. * Absa Group for their foundational and continued support of the Chair of Data Science at the University of Pretoria. * The Department of Computer Science University of Pretoria, UP Faculty of Engineering, Built Environment and Information Technology, and University of Pretoria for giving us a home where impactful research can thrive. South African Centre for Digital Language Resources, National Science and Technology Forum (NSTF), South32, and Department of Science, Technology and Innovation for this incredible honor and for recognising the critical value of research software. 🔮 Looking Ahead We are excited to build on this momentum through the African Institute for Data Science and Artificial Intelligence (AfriDSAI) @ UP, establishing a platform of excellence and societal impact for researchers across the continent. Let's keep building! 🚀 #NSTFawards2026 #NLP #AfricanNLP #OpenSource #DataScienceForSocialImpact #UniversityOfPretoria #ResearchSoftware #LinguisticEquity

Congratulations to the TextAugment Team, recipients of the prestigious NSTF-SADiLaR Research Software Award: Human Language Technologies (HLT) at the 2025/2026 NSTF-South32 Awards! Sponsored by the South African Centre for Digital Language Resources (SADiLaR), this award recognises the TextAugment Team, led by Prof Vukosi Marivate—Professor of Computer Science, Director of the African Institute for Data Science and Artificial Intelligence (AfriDSAI) @ UP at the University of Pretoria. The team's groundbreaking work in developing TextAugment, an open-source library, is helping to overcome data scarcity in Human Language Technologies by generating synthetic training data for under-resourced languages such as isiZulu and Sesotho,. enabling robust AI models and advancing linguistic equity globally. Their innovation is enabling the development of more robust AI models, advancing linguistic equity, and ensuring that African languages are better represented in the future of artificial intelligence. Join us LIVE for the NSTF-South32 Awards Gala Event: https://lnkd.in/dnMjF2i8 South32 Department of Science, Technology and Innovation  #NSTFSouth32Awards #NSTFawards2026 #SADiLaR #ResearchSoftwareAward #HumanLanguageTechnologies #ArtificialIntelligence #AI #DataScience #TextAugment #OpenSource #AfricanLanguages #IsiZulu #Sesotho #LinguisticEquity #SETInnovation #ScienceForSociety #UniversityOfPretoria #SouthAfricanScience #Innovation 

AfriDSAI Fellow Spolight The barrier to better healthcare isn't a lack of research; it is the inability to translate discoveries into clinical practice, policy, and education. Dr. Deborah Oluwadele is bridging this gap by treating clinical entrepreneurship as a distinct discipline. She is equipping clinicians to become systems thinkers who can navigate the complexities of health infrastructure and institutional change. Through platforms like MEDELMETRICX, DOKI, and CLEENET, she focuses on knowledge mobilisation — ensuring that AI-driven innovations in telemedicine and medical education actually scale to deliver public health benefits. She is currently building a pipeline to produce 200 clinical innovators by 2030. As an Emerging Faculty Fellow, she integrates translational medicine with implementation science to ensure African health systems are not just adopting technology, but shaping its trajectory. Her work is a cornerstone of AfriDSAI's mission to turn research into measurable, local impact. "Innovation should not end with discovery. Its true value lies in reaching the people whose lives it is meant to improve." Read the full spotlight → https://lnkd.in/ekPUumpx University of Pretoria, UP Faculty of Engineering, Built Environment and Information Technology, School of Information Technology #AfriDSAI #DigitalHealth #HealthSystems #AIforAfrica

This afternoon I had the pleasure of joining the webinar "Who Does the Algorithm Hear? African Languages, AI Media Platforms, and the Fight for Linguistic Justice," hosted by the Journalism & Media Studies stream of the South African Communication Association (SACOMM) and Wits University (the University of the Witwatersrand), South Africa. It was an engaging discussion on a question that deserves far more attention: whose languages are represented, searchable, and ultimately remembered in digital spaces? The webinar examined how AI-driven media platforms continue to privilege some languages while marginalizing others. These are not simply technical shortcomings. They have real implications for visibility, participation, and the preservation of linguistic and cultural heritage. One point that resonated with me relates to research I have recently begun on music discovery platforms. During the webinar, I shared an observation that many songs in local and minority languages are simply not recognized by popular music identification apps. If a song cannot be identified, indexed, or recommended, it becomes far less visible within today's digital ecosystem. This raises an important question: as more of our cultural memory is mediated through algorithms, what happens to the music these systems fail to hear? The issue extends well beyond convenience. It touches on discoverability, preservation, and the long-term visibility of languages and cultures that remain underrepresented online. As we increasingly rely on AI and digital platforms to organize and retrieve information, algorithmic recognition itself becomes part of cultural preservation. Many thanks to facilitator Fortune Maswanganyi and the outstanding panel of Dr. Tsosheletso Chidi, Trust Matsilele, Unathi Koboka, and Dr. Keaobaka S. for a thought-provoking discussion that sparked several new research ideas. I look forward to following this conversation as AI, language, and digital culture continue to intersect. #ArtificialIntelligence #AfricanLanguages #LinguisticJustice #DigitalCulture #CulturalMemory #AIandMedia #MediaStudies #CommunicationResearch #DigitalHeritage #SACOMM2026 #DigitalEquity

Just finished giving this keynote presentation and what an engaging audience! Thank you to Deep Learning IndabaX South Africa for the invitation and for organising yet another superlative IndabaX South Africa. Thank you to all the attendees and a special “I-love-to-see-it” to those attendees who engaged me with very thoughtful questions. I believe that together we can turn “all hands on deck AI governance” from a question to a statement both in Africa and globally. Data Science Law Lab African Institute for Data Science and Artificial Intelligence (AfriDSAI) @ UP University Of Pretoria, Faculty of Law