Dipartimento di Ingegneria informatica, automatica e gestionale

Titolo del seminario

Visual Analytics for eXplainable Deep Learning

Abstract

Artificial Intelligence, in particular Deep Learning (DL), has become pervasive in many domains in recent years. To help humans to understand the reasoning behind decisions or predictions made by the DL models and improve their trustability, the research field of eXplainable Deep Learning (XDL) has emerged. Visual Analytics (VA), defined as the science of analytical reasoning supported by interactive visual interfaces, is a consolidated research discipline effectively applied to many different domains. During the last five years, Visual Analytics has provided several solutions supporting the emerging eXplainable Deep Learning field. This seminar aims to present the basic notions of Visual Analytics and Deep Learning, and then discuss how these two worlds have been merged to support the emerging eXplainable Deep Learning, showing the current state-of-the-art of the field. 

Short Bio

Graziano Blasilli obtained a PhD in Engineering in Computer Science from Sapienza Università di Roma in 2022, where is currently a post-doctoral researcher. His research interests include Visual Analytics applied to the cybersecurity domain and the eXplainable Artificial Intelligence field. He is part of the A.WA.RE Research Group (http://aware.diag.uniroma1.it/) led by Giuseppe Santucci.