Lorenzo Capelli

PhD Student

Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi"

Teaching tutor

Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi"

Academic discipline: ING-INF/01 Electronic Engineering

Research

Explainable AI (XAI): Visualization of the network’s decision-making process through peepholes and conceptograms.
Trustworthy AI: Verification of the reliability of neural model outputs by analyzing the statistical patterns of intermediate activations. This includes verification using data considered anomalous for the reference models.
Defense against attacks: Analysis of intermediate activations to determine whether the input generated by the network is harmless or constitutes an attack.