Research
My academic research is driven by two primary objectives: developing interpretable machine learning models and designing label-efficient learning algorithms. In an era where AI systems are increasingly employed in sensitive domains, the imperative for explainability is paramount. My work focuses on creating machine learning systems that are not only effective but also transparent and accountable. This involves ensuring these systems are characterized by their accuracy, fairness, and transparency, making them more reliable and trustworthy.
Simultaneously, I address a significant challenge in the field of machine learning: the heavy reliance on large volumes of labeled data. Acquiring such data is often expensive and time-consuming. My research endeavors to mitigate this dependency by developing methods that enable machines to achieve optimal performance with minimal labeled data. This approach is particularly relevant in scenarios where data labeling is impractical or infeasible, making machine learning applications more accessible and efficient.
Publications
- Abdurahman Ali Mohammed, Catherine Fonder, Donald S. Sakaguchi, Wallapak Tavanapong, Surya K. Mallapragada, and Azeez Idris. 2023. IDCIA: Immunocytochemistry Dataset for Cellular Image Analysis. In Proceedings of the 14th ACM Multimedia Systems Conference (MMSys ’23). Paper
- A. Idris, A. A. Mohammed, and S. Fanijo, “Stronger is not better: Better Augmentations in Contrastive Learning for Medical Image Segmentation,” NeurIPS Black in AI workshop, Nov-2022. Poster
