Vincent Abbott

Vincent Abbott is a researcher at the Australian National University (ANU). He completed a B.S. majoring in Computer Science with a minor in Physics from ANU and wrote an honors thesis on “Robust Diagrams for Deep Learning Architectures: Practice and Theory”. His research specializes in diagrammatic category theory and its application to machine learning. His diagrams allow for deep learning systems to be comprehensively expressed as mathematical structures, and allow for clear communication. He has worked with Gioele Zardini on applied category theory projects developing the mathematics underpinning general diagrams and using them to show that the atypical composition of negative information can be considered in a traditional category theoretic manner. Currently, he is working on applying diagrams to guide optimal implementations of algorithms and engages in outreach, using his diagrams to make complex deep learning models and their implementation accessible.