Abdullah

I'm a postdoctoral fellow at Hamad Bin Khalifa University in Doha, Qatar. I build machine learning methods for signals and images that come from the human body and from satellites.

Much of my published work asks a narrow question about EEG: which electrodes actually carry the information? I have answered it with graph-based channel selection, adaptive thresholds and reinforcement learning. I have also built graph neural networks for hyperscanning, where several people are recorded at once.

My current work moves in two directions. In quantum machine learning, I study when a quantum layer earns its place inside a hybrid model, and I use the geometry of the circuit's tangent space to explain barren plateaus. In medical imaging, I review how deep learning corrects motion in PET scans and how well those methods are validated.