UoA ML Seminar: Lia Morra – Neuro-Symbolic AI Advancements in Computer Vision

Short Bio

Lia Morra is Assistant Professor (Tenure Track) with the Department of Control and Computer Engineering, Politecnico di Torino, Italy. She received the M.Sc. and PhD degree in Computer Engineering from Politecnico di Torino, Italy, in 2002 and 2006, respectively. From 2006 to 2017, she was researcher at im3D (Turin, Italy), where she served as Chief Scientific Officer from 2014 to 2017 developing artificial intelligence systems for medical image interpretation. She is an IEEE Senior Member and a member of the AAPM Computer Aided Image Analysis Subcommittee (CADSC). Her main research interests are deep learning, image interpretation, neuro-symbolic artificial intelligence and medical image analysis.

Abstract

Neuro-Symbolic AI (NeSy) techniques are a class of artificial intelligence techniques that focus on integrating symbolic knowledge representation and reasoning within machine learning, and deep learning in particular. This hybrid paradigm holds promise for solving intricate real-world problems involving complex reasoning, as well as designing machine learning systems that can incorporate prior knowledge and learn to satisfy hard and soft constraints. This talk will provide an overview of the current state of NeSy techniques, through the lenses of a specific NeSy framework called Logic Tensor Networks and its applications in computer vision.