Four PhD scholars of the Responsible AI Lab (RAIL), Albert Dede, Amina Salifu, Matthew Cobbinah, and Julius Adinkra, have successfully defended their doctoral research. Their work exemplifies RAIL’s objective to innovatively, ethically, and impactfully advance artificial intelligence for society.
Albert’s research tackles the complex challenges of analysing gigapixel-resolution images, with direct applications in medical diagnostics. By employing deep learning and weakly supervised methods, he developed innovative techniques like wavelet-based feature extraction to process these massive images efficiently. His work, including a seminal systematic review in the field, has been published in leading journals like Engineering Reports.
Beyond his research, Albert is an accomplished reviewer and educator, having served as a Teaching Assistant at KNUST. He credits RAIL for providing crucial financial support, opportunities to attend international conferences, and a collaborative environment that fostered his growth into a well-rounded AI researcher.
Amina’s work intersects Natural Language Processing (NLP) and Machine Learning (ML). She focused on developing deep learning models to detect native and non-native English speakers, with a special emphasis on identifying Ghanaian English accents. Her research lays the groundwork for more inclusive and accurate speech recognition technologies tailored to local contexts.
Her excellence has been recognised internationally, winning awards at the Deep Learning Indaba and the Ghana Data Science Summit. RAIL’s support, including funding and access to a global network of experts, enabled her to present her work on stages like the Women in Machine Learning Workshop in Canada, significantly enriching her academic journey.
Matthew’s research addresses critical challenges in generative AI, specifically using Generative Adversarial Networks (GANs) for unpaired medical image translation (e.g., converting CT scans to MRIs). His novel architectures, DeCGAN and Attn-DeCGAN, are designed to overcome common issues like mode collapse, resulting in synthetic images with superior anatomical fidelity validated by expert radiologists.
His work has profound clinical relevance, potentially providing MRI-like insights from CT scans for patients in resource-limited settings, including Ghana. As a RAIL scholar, Matthew received mentorship and international exposure that deepened his commitment to developing ethically sound and socially responsible AI for healthcare.
Julius is tackling the critical challenge of rural electrification in Africa. His research leverages deep learning and stochastic modelling to create accurate frameworks for forecasting electricity demand in remote, underserved communities. By providing data-informed strategies, his work helps prevent inefficient grid planning, enabling sustainable and equitable energy access.
Published in Renewable and Sustainable Energy Reviews, his research can potentially transform energy policy across the continent. RAIL provided the supportive environment and sponsorship instrumental to his success, including his award-winning poster presentation at the Ghana Data Science Summit. Julius’s work directly contributes to achieving UN Sustainable Development Goal 7.
The Responsible AI Lab congratulates Albert, Amina, Matthew, and Julius on their outstanding achievements. We are proud to have supported their journeys and are excited to see how their future work will continue to push the boundaries of responsible artificial intelligence.