Eva Feng
Interests
My research interests lie in the broad area of artificial intelligence, more specifically in the intersection between machine learning and symbolic reasoning. I am deeply drawn to the curious connections between logic, neural networks, and other knowledge representation formalisms. These forces have driven me to study the correspondence between the expressive power of graph neural networks and logic.
Biography
I obtained my Bachelor's and Master's degrees in Mathematics from the University of Waterloo. For my Bachelor's degree, I double majored in Computer Science as well as Combinatorics and Optimisation and graduated with distinction. I have now moved to the University of Oxford for my doctoral studies. I am fortunately supervised by Prof. Bernardo Cuenca Grau, Prof. Ian Horrocks, and Dr. Przemysław Wałęga. My Dphil is generously supported by a Google DeepMind Scholarship. My Dphil project focuses on the logic of bounded graph neural networks. I studied the expressive power of bounded graph neural networks, in particular, the correspondence between bounded graph neural networks and fragments of first-order logic. I am also interested in the explainability and verification of (bounded) graph neural networks. During my undergraduate and master's studies, I worked on database query optimisation and automated reasoning. I was fortunately supervised by Prof. David Toman at the University of Waterloo (great fun with the winter snow and Mr. Goose et al.).