Join Us for an Exciting Seminar on Designing Automatic Speech Recognition Systems that perform well and Scale on African Low-Resource Languages.! - Dr Hope Mogale

s🌍 Topic: Designing Automatic Speech Recognition Systems that perform well and Scale on African Low-Resource Languages.
🎤 Speaker: Dr. Hope Mogale
📅 Date: 26 March 2025
⏰ Time: 14:00 - 15:15 SAST

RSVP/Sign UP here https://forms.gle/hXdtkFA3bzcTKCwM8

🔎 About the Speaker

Dr. Hope Mogale is a South African-born computer scientist with a PhD in Computer Science from North-West University, obtained in 2019. His research spans across Machine Learning (ML), Artificial Intelligence (AI), and Natural Language Processing (NLP), with a particular focus on developing technologies for low-resource African languages. He specializes in Large Language Models (LLMs), Generative AI, High-Performance Computing (HPC), and advanced computing architectures like ASICs and FPGAs, as well as Parallel Deep Neural Networks. Dr. Mogale is currently a Senior Member of the DSFSI Lab at the University of Pretoria and a Postdoctoral Fellow funded by Microsoft Research and the Department of Research and Innovation (DRI) at the University of Pretoria. His current work addresses challenges in Automatic Speech Recognition (ASR) for African languages, where he focuses on developing robust ASR systems that perform well on low-resource datasets. His research emphasizes overcoming the unique linguistic and computational challenges faced by African languages. Through his work on NLP and self-supervised learning, he is exploring innovative ways to bridge the digital language divide and improve access to AI-driven speech technologies across the continent.

Outside of his research, Dr. Mogale enjoys music, designing, hiking, tennis, basketball, travelling, chess, F1 and a bit of gaming when he gets a chance.

[

](https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F36787846-95ad-4e88-a000-5ba10057b534_1200x628.jpeg)

📖 What to Expect

Automatic Speech Recognition (ASR) systems have seen significant advancements globally; however, African low-resource languages remain largely underserved due to data scarcity, linguistic diversity, language exclusivity, model limitations, and computational constraints. Despite the growing interest in speech technology, many African languages remain underrepresented in existing datasets and ASR models, limiting their accessibility and usability in real-world applications. This seminar aims to equip young researchers and ASR practitioners with both the theoretical foundations and practical methodologies necessary for designing high-performance ASR systems tailored to low-resource African languages. We will introduce a fully integrated ASR research suite capable of detecting low-resource dataset formats, training and refining models, and identifying synthetic speech. Through hands-on demonstrations, we will explore critical technical aspects, including dataset preprocessing, model adaptation techniques, and error recovery mechanisms optimized for African linguistic structures. Special emphasis will be placed on the challenges of working with limited annotated data, the impact of phonetic variability, and strategies to improve ASR robustness in low-resource settings and applications on real world problems. By addressing both fundamental principles and advanced techniques, this seminar seeks to advance African speech technology research and innovation. Attendees will gain insights into overcoming key bottlenecks in low-resource language ASR development, ultimately contributing to the creation of more inclusive and effective speech technologies for African languages.

Join us as we explore:
✅ Challenges in Developing ASR systems for African Low-Resource Languages
✅ Innovative ASR Research Framework for Low-Resource Settings
✅ Advancing African Speech Technology Through Practical Solutions

📍 Hosted by: The Data Science for Social Impact Research Group, Department of Computer Science, University of Pretoria

📩 How to Join
A Zoom link will be sent to registered attendees a day before the seminar. Don’t miss this opportunity to engage with cutting-edge research shaping the future of African language technology!

🔗 Save the Date & Stay Tuned!