Yasmine Djebrouni - Accessit au prix de thèse du GDR RSD 2025 | Laboratoire d'InfoRmatique en Image et Systèmes d'information

Yasmine Djebrouni – Honorable Mention for the GDR RSD 2025 Thesis Prize

The thesis is distinguished by an impressive experimental campaign and high-level publications. The jury praised the very solid methodology and the analytical rigor of the work on emerging topics at the frontier between systems research and machine learning.

The thesis of Yasmine Djebrouni, entitled “Characterization and Optimization of Distributed Machine Learning Systems: Towards a Multi-Objective Approach,” carried out at LIG and LIRIS under the supervision of Vania Marangozova and Sara Bouchenak, explores on the one hand the impact of system configurations on the execution of different machine learning models, and on the other hand proposes a modular technique for bias mitigation. The thesis is distinguished by an impressive experimental campaign and high-level publications. The jury praised the very solid methodology and the analytical rigor of the work on emerging topics at the frontier between systems research and machine learning.