Understanding Perceptive Deep Learning and Generative AI in Automotive Applications

This is an external activity approved by the PhD Board, which implies that:

• it allows attendees only to request credit validation;
• it does not allow attendees to accumulate attendance hours (applicable to cycle 38 and 39).

No credit request form or Supervisor's approval is required. Credits will be registered automatically.

ABSTRACT

Today, modern Artificial Intelligence techniques are revolutionizing today's society and science, offering technological solutions capable of carrying out tasks or solving problems in an efficient and robust way. From the industrial field to the healthcare and automotive areas, modern deep learning techniques allow the development of cutting-edge technologies of the highest scientific value. Specifically, in this seminar we will focus on modern perceptual, explainable and generative AI techniques. The theory of perceptive, explainable and generative AI as a bio-inspired system, allows us to emulate neuronal functions of human brain which transfer characteristics of efficiency, robustness as well as high level performance to the developed solutions. During the seminar, after an introduction to the theory of perceptive, explainable and generative AI, we will show a series of applications in the automotive and industrial fields developed as part of STMicroelectronics' R&D activities in the “AI world”.

Specifically, we will show the status of the EU funded Project “R-PODID” in which the aforementioned perceptive deep learning and generative AI methodologies are being applied for monitoring the state of degradation of power systems and latest generation electric engines. Some preliminary results obtained thanks to the collaboration between STMicroelectronics and the University of Bologna, both partners of the R-PODID project, will be discussed. The seminar will conclude with an overview of the ST's MCU hardware solutions for optimal hosting of the delivered AI-based systems.

BIO

Francesco Rundo is a Senior Technical Staff Member and R&D Manager of STMicroelectronics, Catania site. Currently, he leads the Artificial Intelligence Team at the QMT R&D Power and Discretes Division of STMicroelectronics, Catania. He received his degree in Computer Science Engineering and Ph.D. in “Applied Mathematics for Technology” from the University of Catania. In his role as R&D Manager of the AI team of STMicroelectronics, Catania, he has established several scientific collaborations with various research institutes and universities at national and international level. He has co-authored more than 105 scientific contributions in international journals, conference proceedings and book chapters, having a SCOPUS h-index score of 22. He serves as Reviewer, Associate Editor and Guest Editor of several scientific journals in the field of Artificial Intelligence applied to automotive and industrial applications. He also serves as a teacher on Deep Learning topics applied to the industrial and automotive world, at the PhD course in computer science at the University of Catania. He is a member of the Computer Science Ph.D. Committee at the Department of Mathematics and Computer Science of the University of Catania. He is also a scientific board member of the National Ph.D. Course of Artificial Intelligence. He is also co-inventor of several international patents filed in EU and USA. He is a member of several international conference program committees as well as Program and Co-chair on special section and workshop organized in the framework of key-conference in AI (ICCV, ECCV, CVPR, ICASSP, etc..). He serves as Project Leader, WP Leader and researcher in several funded EU Scientific Projects in the field of AI and Deep Learning for automotive and industrial applications (NEUROKIT2E, ARCHIMEDES, R-PODID, EdgeAI, EdgeAI-Trust, etc..). His main research interests include advanced bio-inspired models, advanced and perceptual deep learning, embedded systems for deep learning algorithms, advanced deep learning and mathematical modeling for automotive and industrial applications.