KNUST student explores AI for gloss-free Ghanaian Sign Language translation

Date: 2026-10-17
news-banner

By:   Nana Appiah Acquaye

Lamin Abubakar, a third-year Mechanical Engineering student at Kwame Nkrumah University of Science and Technology (KNUST), Kumasi, has completed an internship with the Responsible AI Lab (RAIL), where he explored the use of artificial intelligence to translate Ghanaian Sign Language directly into natural-language text.

His internship project focused on gloss-free Ghanaian Sign Language translation using Vision Transformers and Large Language Models. He explored how pretrained models such as ViViT and Gemma could be adapted to translate sign language videos directly into natural-language text.

Abubakar concluded his internship by presenting his project to the RAIL team, sharing his approach and findings with colleagues. The feedback and comments he received helped him reflect on his methods and identify areas for further improvement.

Passionate about exploring how AI can solve real-world problems, Abubakar is working towards growing his knowledge and skills as an AI researcher.

He described his time at RAIL as both rewarding and educational. Beyond the technical knowledge he gained, he enjoyed collaborating with colleagues, exchanging ideas, discussing challenges and learning from their different perspectives.

Presenting his work and receiving constructive feedback also helped him grow in confidence and develop a deeper appreciation for collaborative AI research.

RAIL congratulated Abubakar on completing his internship and wished him continued success as he advances his journey in AI research.

 

Leave Your Comments