About me

Hi! I am Francesco Diana, a third year PhD student at Inria in the COATI team, working under the supervision of Dr. Chuan Xu and Dr. Giovanni Neglia.

My research focuses on uncovering privacy vulnerabilities in federated learning by designing novel privacy attacks.

News

  • [April 2026] New preprint! No More Guessing: a Verifiable Gradient Inversion Attack in Federated Learning is available
    [paper]
  • [January 2026] I will be participating to the PEPR Cybersécurité Winter School in Aussois from 26th to 30th January

  • [May 2025] Our paper Cutting Through Privacy: A Hyperplane-Based Data Reconstruction Attack in Federated Learning has been accepted at UAI 2025