Nur Haznirah Hazman, a Master’s student under the Faculty of Data Science and Computing (FSDK) and a researcher at AIBIG. She holds a Bachelor of Computer Science (Bioinformatics) from Universiti Teknologi Malaysia (UTM), Skudai, Johor and is currently focusing her research on Artificial Intelligence in Bioinformatics and Biomedical Data Science.

Her research interests include developing deep learning algorithms and auto-encoders to tackle data scarcity challenges in biomarker discovery for medical research. She has also been actively involved in the Polar Bytes project, which focuses on data centralization and visualization for polar research. Her contributions to this project led to the successful publication of a research paper titled “PolarBytes: Advancing Polar Research with a Centralized Open-Source Data Sharing Platform”, submitted to Elsevier – Environmental Modelling and Software Journal. This initiative compiles data from reputable online repositories, presenting it in an accessible and structured format. The Polar Bytes project received research funding from the Malaysia Antarctica Sultan Mizan Research Foundation (YPASM).

In addition to her work on Polar Bytes, Nur Haznirah is actively conducting research in cancer studies, addressing a key challenge in medical AI—limited datasets for disease detection and analysis. One of her primary focuses is glioblastoma, an aggressive form of brain cancer. AI models often struggle with insufficient medical data, making early diagnosis and treatment predictions difficult.

To address this issue, she is utilizing Generative Adversarial Networks (GANs) to generate high-quality synthetic glioblastoma data. However, traditional GANs have limitations such as mode collapse, where they fail to represent the full diversity of rare mutations. To enhance the reliability of synthetic medical data, she is exploring Hybrid AI Models, including:

• GAN-VAE (Generative Adversarial Networks - Variational Autoencoders): – A combination of GANs' ability to generate realistic data and VAEs' capability to capture underlying structures.
• Diffusion-GAN: Leveraging diffusion models to produce high-quality synthetic data, addressing variations that traditional GANs may miss.

By integrating these advanced AI techniques, her research aims to bridge data gaps, enhance AI-driven medical diagnostics, and contribute to advancements in brain cancer research. Through her expertise and dedication, Nur Haznirah is making significant contributions to AI in healthcare, reinforcing AIBIG’s mission to drive multidisciplinary AI research with real-world impact.