Prof. Fabio Roli
University of Cagliari, Italy
The Rise of Adversarial Patterns in the Cyber-physical Space
Abstract: Pattern recognition systems are currently used for security applications like biometric identity recognition and spam filtering that strongly differ from the traditional ones. The difference lies on the fact that in these applications an intelligent adversary can purposely create adversarial patterns to evade the system or mislead its training. We advocate that the rise of the cyber-physical space will rise the amount of adversarial patterns. As pattern recognition systems will become more and more pervasive, the economic incentives of cyber criminals to attack them will increase. But also ordinary people will be always more motivated to evade pattern recognition systems, to preserve their privacy. In this talk, I discuss the fundamentals of adversarial classification from the perspective of a designer of pattern recognition systems, with examples from security applications like spam filtering and biometric recognition in the cyber-physical space.
Biographical note: Fabio Roli is professor of computer engineering and director of the PRA Lab. His research over the past twenty years addressed the design of pattern recognition systems in the context of real applications. He played a leading role for the research field of multiple classifier systems. He is a prolific author with a value of h-index of 42 according to Google Scholar. He is Fellow of the IEEE and Fellow of the IAPR.