Romain Xu-Darme

Romain Xu-Darme

Short biography

After graduating from ENSIMAG in 2008, I have worked for 12 years in the field of cybersecurity for various companies and public institutions, with a specialization in high-performance computing and embedded security.
I have always been concerned about the safety and security challenges raised by ML-based software and I have decided to start a PhD in the field of interpretable AI in 2020 in order to help develop the methods and tools to bring more trust in those systems.

Thesis subject

Learning logical rules for the formal proof of deep neural networks: application to object classification in the context of autonomous vehicles.

  • Director: Georges QUENOT (LIG/MRIM)
  • Co-supervisor: Marie-Christine ROUSSET (LIG/SLIDE)
  • Co-supervisor: Zakaria CHIHANI (CEA-LIST)


LIG / MRIM team