David Herron

I’m a PhD student of Artificial Intelligence. My research focuses on visual relationship detection within images using hybrid, neurosymbolic systems where the symbolic components are semantic web knowledge graphs and OWL ontologies.

The macro context of my research involves blending the neural (connectionist) and symbolic traditions of AI that, throughout most of AI’s history, have operated as tribal academic silos. The idea is to explore ways of combining deep learning (neural networks) with symbolic knowledge representation and reasoning (in the form of OWL ontologies, knowledge graphs and rules) in order to get the best of both worlds and demonstrate that hybrid, neural-symbolic AI systems can outperform AI systems that rely on neural AI (statistical, inductive, deep learning) alone. An underlying motivation is to help show that semantic web technologies, ontologies and knowledge graphs are cool and can make valuable contributions to AI.