We are offering a PhD position in Applied Mathematics and Machine
Learning, focused on the intersection of reinforcement learning,
optimization, and dynamical systems identification.
The project aims to develop reinforcement learning methods for the
automatic discovery of nonlinear dynamical systems from data, combining
ideas from:
reinforcement learning and sequential decision-making
system identification and inverse problems
optimization and control theory
representation learning for structured dynamical models
The goal is to design algorithms that can learn governing equations or
compact dynamical representations from observations, with a focus on
scalability, robustness, and theoretical guarantees. Applications will
be used as motivating benchmarks, but the core emphasis is
methodological: building general-purpose frameworks for data-driven
discovery of dynamical systems.
This position is suited for candidates interested in theoretical and
algorithmic aspects of modern ML, particularly those working at the
interface of reinforcement learning, optimization, and mathematical
modeling.
Location: Montpellier, France
Deadline: May 4
Applications must be submitted via the doctoral school ED I2S portal
(Université de Montpellier):
https://edi2s.umontpellier.fr
→ PhD offers
→ Statistics and Data Science
→ LPHI – Reinforcement learning for automatic model discovery
Kind regards,
Ovidiu Radulescu