Data Science for Social Impact

📢 Exciting News! DSFSI is at #EMNLP2024! We're thrilled to announce that two of our papers are being presented, highlighting our commitment to advancing #AfricanNLP and pushing the boundaries of #LowResourceLanguages in NLP. 📄 Paper 1: "From N-grams to Pre-trained Multilingual Models For Language Identification" Authors: Thapelo Sindane, Vukosi Marivate Summary: This work explores N-gram models and pre-trained multilingual models (like mBERT, RemBERT, XLM-r, and Afri-centric models such as AfriBERTa, Afro-XLMr, AfroLM, and Serengeti) for Language Identification (LID) across 11 South African languages. Through extensive experiments, we demonstrate that model choice and dataset selection are crucial, revealing Serengeti as a superior performer and introducing a lightweight BERT-based LID model (za_BERT_lid) that shows promising results. 📎 Read the full paper - https://lnkd.in/dr6AeEyn 📄 Paper 2: "Correcting FLORES Evaluation Dataset for Four African Languages" Authors: Idris Abdulmumin, Sthembiso Mkhwanazi, Mahlatse Mbooi, Shamsuddeen Hassan Muhammad, PhD, Ibrahim Said Ahmad, Neo Putini, Mathebula Miehleketo, Matimba Shingange, Tajuddeen Gwadabe, Vukosi Marivate Sumamry: This paper outlines the corrections to the FLORES dataset for Hausa, Northern Sotho, Xitsonga, and isiZulu. Through a careful review by native speakers, we identified and corrected inaccuracies, enhancing the data’s reliability and improving the integrity of NLP evaluations, especially in machine translation. We emphasize the importance of native speakers' involvement for linguistic accuracy in future NLP endeavors. 📎 Read the full paper https://lnkd.in/dvxmVZmG Thapelo Sindane, an incoming PhD student (2025) and recent Masters graduate at DSFSI, will be at #EMNLP all week. Feel free to reach out to him to learn more about our work and our commitment to elevating African languages in NLP! 🌍✨ #EMNLP2024 #LanguageTechnology #MachineTranslation #AIforAfrica #DataScienceforSocialImpact

The deadline has been extended to 18 November 2024. Spread the news and share with your networks.

Together Data Science Law Lab 🚀

📢💼 We Are Hiring! Attention University of Pretoria full-time 3rd year, honours/4th year or masters students! DSFSI is seeking a part-time Research assistant and/or Data engineer who will be working (10-20 hrs/week), starting January 2025. You will be involved with several innovative research projects currently underway within the group. Check out our past projects at https://buff.ly/3Avrg4J Ideal candidates should have or be willing to develop skills in machine learning, software development for pipelines, and data analysis. Individuals with diverse backgrounds in computational, mathematical, statistical, or engineering fields are invited to apply. Apply now to join our team and gain hands-on experience. Application deadline: 20 November 2024 For more info and to apply:

📢💼 We Are Hiring! Attention University of Pretoria full-time 3rd year, honours/4th year or masters students! DSFSI is seeking a part-time Research assistant and/or Data engineer who will be working (10-20 hrs/week), starting January 2025. You will be involved with several innovative research projects currently underway within the group. Check out our past projects at https://buff.ly/3Avrg4J Ideal candidates should have or be willing to develop skills in machine learning, software development for pipelines, and data analysis. Individuals with diverse backgrounds in computational, mathematical, statistical, or engineering fields are invited to apply. Apply now to join our team and gain hands-on experience. Application deadline: 20 November 2024 For more info and to apply:

We keep building upward 🚀

Last month, The Center for Digital Humanities at Princeton University kicked off the African Languages in the Age of AI Speaker Series with a full house for Vukosi Marivate's talk, "A New Agenda for African Languages x AI: Everything, Everywhere, All At Once". ➡ 📹 The recording is now available online at: https://lnkd.in/exd4qT-V. In his talk, Professor Marivate (Chair of Data Science, Professor of Computer Science, University of Pretoria) discussed the crucial role of community building in developing technologies for African languages in the age of AI and touched upon the unique challenges and opportunities in fostering collaboration for African languages, developing technologies that respect and empower communities, and his vision for the future of technology and community engagements in African languages. Simon Gikandi (Chair of Princeton University's Department of English) served as respondent and moderator of the Q&A. The African Languages in the Age of AI speaker series is co-sponsored and supported by: the Africa World Initiative, Program in African Studies, and the African Humanities Colloquium / Princeton Institute for International and Regional Studies. Photography by Alison Nugent.

Catch up on the insightful #DS4SocietySeminar presented by Moseli Mots'oehli from the University of Hawaii at Manoa. The seminar, "Towards Safer Driving on Southern African Roads Through Assistive and Autonomous Annotation and Driving" focuses on enhancing assistive and autonomous perception software to promote road safety for drivers in Southern Africa. This initiative addresses the urgent need to reduce road accidents, with statistics showing a high rate of traffic incidents in the region. The full video recording is now available: https://lnkd.in/dfbrsYms #UPTuks #DSFSI #RoadSafety

On the 30th of September 2024 the Embassy of Sweden in Pretoria and the Computer Science department at the University of Pretoria UP Library hosted an enlightening AI Workshop. The focus was on Low Resource Languages and Large Language Models in Africa: https://lnkd.in/ddTb6vRS and The role of Academia, Industry and government in AI and R&D: https://lnkd.in/d6AKGibQ. Lesego Makhafola and Vukosi Marivate moderated the sessions. The 1st session focused on developing LLMs for low-resource languages in Africa, focusing on limited data and resource challenges. Emphasis was placed on fostering collaboration to create inclusive AI that supports diverse languages and preserves linguistic diversity. The 2nd session focused on AI's transformative potential in healthcare and education, alongside challenges for small businesses in adopting AI. This session emphasized collaboration between academia, industry, and government to boost AI education and policy support, while addressing ethical and sustainability concerns for inclusive, context-specific AI solutions in Africa. The panel included accomplished individuals such as Vukosi Marivate from DSFSI, Paul dos Santos from AI Sweden, Dr. Chijioke Okorie Okorie from Data Science Law Lab, Nomonde Katlego Khalo from the University of Cape Town, Martin Svensson from AI Sweden, Saidah Nash Carter from Bright Insight Global LLC and Greg Desilla from Melio AI. The discussions were insightful, sparking ideas for the future of AI in Africa. #UPEBIT #EmbassyofSwedenofPretoria #UPTuks UP Faculty of Engineering, Built Environment and Information Technology

🚀 Exciting #INSPIRE FILM Update - Diabetes Management Through Innovation! 🚀 We are excited to share the highlights of our latest research project at Stellenbosch University's Division of Health Systems and Public Health. 🎥🌍 Discover how we are bridging #technology and #healthcare to create more inclusive solutions for global health. 📱🤝 Film Credits: 🎬 Film Production: The Dollie House 🎤 Speakers:  - Prof. Lynn Hendricks - Gabriela Carolus - Dr. Jacki O'Neill - Prof. Simone Titus Dawson - Dr. M. Motloung - Mercy Muchai 👥 Participants: Pricilla Anthony; Keanu Jansen Van Vuuren; M Grantham; Lynn Booysen; Dr. DC Joubert; Karabo Dinelelo; Dr. Lee-Ann Gillion; Monique de Wit; Melissa Kelly; Gadija Claasen; Bridget McNulty; Shiara Pillay; David Titus; Nophiwe Job; Monika Deschodt; Lourentia van Wyk; Priscilla Adams; Jodie Layman-Lemphane; Zisikazi Jardine; Amber Cupido. 🤝 Supported by Prof. Rene' English, the Division of Health Systems and Public Health, Department of Global Health, the INSPIRE COLLAB Team (Mandisa Mireldah Mashaba; Nikki Thomas; Nicala Zeeman; Amina Abdullah), and the collaboration with the Data Science for Social Impact team (Prof Vukosi Marivate and Dr Hope Mogale), Boyd Migisha (Swansea University) and the Microsoft Research Africa, Nairobi Team (Mercy Muchai; Stephanie Nyairo; Najeeb G. Abdulhamid, PhD). #GenerativeAI #HealthcareInnovation #DiabetesCare #AIforGood #DigitalHealth #StellenboschUniversity #Storytelling #CoDesignWorkshop #TechForHealth #AIInHealthcare