Muhammad Umair Danish, PhD

I am a Postdoctoral Fellow in Applied AI at Western University. My research focuses on explainable and interpretable artificial intelligence, reliable temporal learning, physics-guided machine learning, and the evaluation of machine-learning systems in applied settings.

I completed my PhD in Electrical and Computer Engineering at Western University, where my dissertation examined neural networks and their interpretability for building-energy modeling. My work has since expanded across mechanistic interpretability, human-centered explainable AI, physics-guided learning, time-series forecasting, perceptual evaluation of AI-generated images, and research software.

Across these areas, I am interested in a common question: how can machine-learning systems provide evidence that is technically meaningful and useful to people? My research combines model development with careful baselines, ablations, statistical validation, visualization, and human-centered evaluation where appropriate. I currently work under the supervision of Dr. Katarina Grolinger and collaborate with researchers across computing, engineering, health, and music therapy.