I am a Ph.D. candidate in the Dept. of Informatics, Systems and Communication of the University of Milan-Bicocca. My supervisors are prof. Marco Antoniotti and Dr. Alex Graudenzi. From March 2022 I’m doing an internship at IBM Research Zürich supervised by Dr. María Rodríguez Martínez.

 

My PhD is focused on characterising anomalous input instances, including adversarial examples, from the hidden features of deep neural networks. In the first year of my PhD, I focused on defining new techniques to analyse the relationship between data and the learned function of piecewise linear deep neural networks. In my second year, I focused on methods to detect adversarial attacks from the hidden activations of state-of-the-art convolutional neural networks.

 

Keywords: Adversarial examples detection, deep learning, explainable AI.