Perceval Beja-Battais

About me

I’m a 3rd year PhD Student in Applied Mathematics at Centre Borelli (ENS Paris-Saclay) under supervision of Nicolas Vayatis. My PhD is funded through a CIFRE fellowship with Framatome, a world leader in civil nuclear industry. My research sits at the intersection of machine learning and optimal control, with a focus on developing algorithms that are both faster and more accurate in demanding industrial settings. I am particularly interested in how modern ML frameworks can address the computational challenges that arise in real-world control problems.

All my publications are listed here and on my Google Scholar.


Recent Works

Research Internship Supervision -- Efficient Sampling of Trajectories for Online Finetuning of ML Surrogate Simulation Scheme (Starting May 2026)
Research Internship Supervision -- Uncertainty Quantification for DAE Simulation Schemes (Starting Mar 2026)
Towards Safe Industrial Control using Machine Learning Surrogates -- Ongoing work, to be published soon! We leverage a fast ML surrogate integration scheme for Optimal Control Problems, especially long-horizon Nonlinear Model Predictive Control.
Leveraging Machine Learning to accelerate Differential Algebraic Equations simulation algorithms -- See Paper
This paper develops a ML surrogate simulation scheme for fast integration of Differential Algebraic Equations modeling a nuclear reactor core. illustration
Designing practical improvements on a global optimization algorithm -- See Paper
This paper develops two major improvements on the global optimization LIPO from Malherbe & Vayatis, 2017: an empirical stopping criterion and a decaying exploration rate. illustration
A theoretical review of AdaBoost -- See Paper Unifying the views of AdaBoost, in order to better understand its dynamics.

Office 3S28
ENS Paris-Saclay
Gif-sur-Yvette, France
perceval.beja-battais [at] ens-paris-saclay [dot] fr
🎓
🍝 Webring des délices 🍜