PhD Student · University of Toronto

Yifan Qu

Graph Neural Networks & Numerical Methods
for Scientific Computing

Scroll
Yifan Qu

I am a PhD student at the University of Toronto, advised by Prof. Christina Christara. My research sits at the intersection of graph neural networks, numerical methods, and scientific computing — with a focus on developing efficient computational methods for PDE-based models, numerical algorithms, and image processing. I am a member of the Numerical Analysis Group at UofT CS.

I received my BSc Honours in Applied Mathematics and Computational Mathematics from the University of Waterloo, where I was advised by Prof. Hans De Sterck and Prof. Andrea Scott.

Outside of research, I enjoy photography, piano, ballet, and figure skating, and am always eager to explore interdisciplinary collaborations.

UofT CS GNNs PDEs Numerical Methods Scientific Computing
Graph Learning

River Ice Classification with GNNs

Developed a learnable edge-weight GNN for SAR image classification, achieving a 10% accuracy improvement over CNNs.

View paper
Deep Learning · PDEs

PDE-Based Graph Neural Networks

Designed deep GNN architectures using first-order PDEs to mitigate over-smoothing and enhance learning in deep networks.

View preprint
Image Processing

Underwater Image Dehazing

Improved the Dark Channel Prior (DCP) method, reducing mean squared error by 10%.

View paper
Numerical Finance · In Progress

American Option Pricing via Free Boundary Methods

Numerical approximation of the early exercise boundary of American options using PDE-based free boundary formulations. Joint work with Prof. Christina Christara.

See notes

Feel free to reach out — I'm happy to chat about research, collaborations, or anything interesting.