Mahbaneh E
- Research Program Mentor
PhD at University of Pittsburgh
Expertise
Artificial Intelligence, Machine Learning, Computer Vision, NLP, Gren AI, LLMs
Bio
I am deeply passionate about the intersection of artificial intelligence, healthcare, and scientific discovery, with a strong academic foundation in machine learning, computer vision, and neuroimaging. My research has focused on developing deep learning methods for medical image harmonization, predictive modeling of clinical and genomic data, and knowledge-augmented rule learning. Recently, I’ve expanded my work to include generative AI and large language models (LLMs), exploring how they can power multi-modal systems and enhance data-driven decision-making in biomedical research. Outside of work, I find inspiration in creative expression, whether through literature, writing, or exploring nature on long hikes. I enjoy learning across disciplines and thinking about how technology and human insight can come together to solve meaningful problems. Curiosity drives me—whether I’m diving into a new machine learning framework or getting lost in a good book.Project ideas
Can We Make Brain Scans from Different Hospitals Look More Similar?
When scientists use brain scans (like MRIs) from different hospitals, the images can look a little different because they were taken with different machines. This can confuse computer programs that are trying to learn patterns, like figuring out who might have a disease based on their brain scan. In this project, you'll explore how we can make these scans look more similar—this process is called harmonization. You’ll learn how to use AI tools like convolutional neural networks (CNNs), which are great at looking at images, and generative AI methods like GANs (generative adversarial networks), which can help transform one type of image to look like another. You’ll also explore domain adaptation methods—these help AI models adjust when the data comes from different places. The goal is to teach the computer to ignore the differences between scanners and focus on what’s really important: the patterns in the brain that relate to health. This is a great way to dive into machine learning, computer vision, and real-world healthcare problems using brain imaging data.