Explainability and Privacy-Preserving Data-Driven Models
Vicenc Torra, Umea University, Sweden
Artificial Intelligence for Biometrics
Vincenzo Piuri, Università degli Studi di Milano, Italy
Explainability and Privacy-Preserving Data-Driven Models
Vicenc Torra
Umea University
Sweden
Brief Bio
Vicenç Torra is Professor on AI at Umeå University (Sweden). He is IEEE and EurAI Fellow, and ISI elected member. He held positions at the Spanish Research Council (IIIA-CSIC), Maynooth University (Ireland), and Skövde University (Sweden). His fields of interests include privacy-preserving machine learning, and approximate reasoning (fuzzy sets and non-additive measures).He has written several books including "Modeling decisions" (with Y. Narukawa, Springer, 2007), "Data Privacy" (Springer, 2017), "Guide to Data Privacy" (Springer, 2022). He is founder and editor of the Transactions on Data Privacy.
Abstract
Explainable AI has become a requirement for most intelligent systems. Automated decisions need to be explained, and, thus, a collection of tools have been developed to help providing explanations of the models themselves, and on about the decisions taken. Data-driven models naturally strongly relate to the data. Nevertheless, as data is sensitive, privacy preserving machine learning models are a must. For this masking methods have been developed, as well as other approaches to directly build privacy-preserving machine learning models from unprotected data. In this talk we will discuss the problem of building explanations for data-driven models when data is sensitive.
Artificial Intelligence for Biometrics
Vincenzo Piuri
Università degli Studi di Milano
Italy
![]()
Brief Bio
Vincenzo Piuri is Full Professor in computer engineering at the University of Milan, Italy (since 2000). He has been Associate Professor at Polytechnic of Milan, Italy and Visiting Professor at the University of Texas at Austin, USA, and visiting researcher at George Mason University, USA. His main research interests are: artificial intelligence, computational intelligence, machine learning, pattern analysis and recognition, intelligent systems, signal and image processing, biometrics, and industrial applications. Original results have been published in 400+ papers in international journals, proceedings of international conferences, books, and book chapters. He is Fellow of the IEEE, Distinguished Scientist of ACM, and Senior Member of INNS. He is IEEE Region 8 Director (2023-24),and has been IEEE Vice President for Technical Activities (2015), IEEE Director, President of the IEEE Systems Council, President of the IEEE Computational Intelligence Society, and Vice President for Education of the IEEE Biometrics Council. He has been Editor-in-Chief of the IEEE Systems Journal (2013-19) and has been Associate Editor of the IEEE Transactions on Computers, the IEEE Transactions on Cloud Computing, the IEEE Transactions on Neural Networks, and the IEEE Transactions on Instrumentation and Measurement. He received the IEEE Instrumentation and Measurement Society Technical Award (2002), the IEEE TAB Hall of Honor (2019), and the Rudolf Kalman Professor Title of the Obuda University, Hungary. He is Honorary Professor at: Obuda University, Hungary; Guangdong University of Petrochemical Technology, China; Northeastern University, China; Muroran Institute of Technology, Japan; Amity University, India; Galgotias University, India; Chandigarh University, India; and BIHER, India.
Abstract
Biometrics concerns the study of automated methods for authenticating an individual or identifying her among many people by measuring one or more physical or behavioral traits of her, like fingerprint, face, iris, voice, and gait. Biometrics was originally used for critical applications, like access control for critical areas and services. Nowadays, the advancements in technology and biometric science allow for creating many more applications and become increasingly pervasive in our daily life. The use of biometrics in the real life often requires very complex signal and image processing and scene analysis, for example encompassing biometric feature extraction, tracking, liveness/anti-spoofing tests, facial expression recognition, and operation in less-constrained conditions. Artificial intelligence (including neural networks, deep learning networks, fuzzy logic, evolutionary computing, and multi-agent systems) have been proved to be extremely useful and effective in addressing all the above kinds of data processing, especially when is difficult to identify a traditional algorithmic approach due to the complexity of the problem. This keynote will review the domain of biometrics and discuss challenges and opportunities offered by artificial intelligence for effectively solving various problems in biometric applications.