Hybrid AI for Knowledge Representation and Model-based Image Understanding

Explore hybrid AI approaches for spatial reasoning and image understanding in this 48-minute lecture by Isabelle Bloch from LIP6/Sorbonne Université and LTCI/Télécom Paris. Delve into the combination of different formalisms and methods that merge knowledge with data to enhance explainability in artificial intelligence. Examine various knowledge representation techniques, including symbolic, qualitative, and semi-qualitative approaches, used to model structural information through graphs, ontologies, logical knowledge bases, and neural networks. Learn how image understanding is framed as a spatial reasoning problem and how hybrid AI contributes to more effective scene and object analysis. Discover practical applications of these concepts in medical imaging, showcasing the benefits of integrating multiple approaches for improved results.