ANR Project OPTIMAL
Optimisation des sources optiques ultra-rapides à large bande à l'aide de l'apprentissage automatique
Optimized on-demand ultrafast and broadband light sources using machine learning
OPTIMAL - ANR-20-CE30-0004
Résumé
OPTIMAL vise à développer des sources lumineuses avancées en appliquant les techniques de l'apprentissage automatique à la conception des sources laser et à l’étude de l'optique non linéaire ultra-rapide. Les objectifs spécifiques sont : (i) concevoir de nouvelles approches pour générer la lumière à large bande avec des propriétés spectrales et temporelles sur mesure en utilisant l'apprentissage automatique pour optimiser les conditions initiales de propagation non linéaire dans des guides photoniques ; (ii) développer de nouvelles sources ultra-rapides basées sur le concept d'oscillation régénérative de Mamyshev utilisant l'apprentissage automatique pour optimiser le contrôle intra-cavité ; (iii) combiner le développement de ces nouvelles sources avec des études fondamentales de dynamique non linéaire, s'inspirant de l'apprentissage automatique pour analyser des données expérimentales, et en interprétant les résultats en termes des analogies avec d’autres systèmes physiques.
Summary
OPTIMAL aims to develop advanced ultrafast and broadband light sources by bringing the power of machine learning into the mainstream of laser design and ultrafast nonlinear optics. Specific objectives include: (i) designing new approaches to generate broadband light with tailored spectral and temporal properties using machine-learning to optimize initial conditions for nonlinear propagation in optical fibre waveguides; (ii) developing new ultrafast sources based on the Mamyshev regenerative oscillation concept using machine-learning to optimize intra-cavity control of amplitude and phase; (iii) combining the development of these novel sources with fundamentally-oriented studies of the underlying nonlinear propagation dynamics, building on analogies such as cavity hydrodynamics and turbulence, and using machine learning to analyse and interpret experimental data.
Consortium
FEMTO-ST, UNIVERSITE MARIE ET LOUIS PASTEUR
ICB, UNIVERSITE BOURGOGNE EUROPE
XLIM, UNIVERSITE DE LIMOGES
Summary of Achievements
The project OPTIMAL ran from XX/XX/2020 to XX/XX/2025, and resulted in a number of significant advances in the field of machine learning applied to ultrafast nonlinear fibre systems.
OPTIMAL has resulted in a number of major advances relating to both the targeted theoretical/numerical and experimental objectives. In particular:
1. Theoretical and numerical studies
- Data-driven discovery of the underlying physical model governing four-wave mixing in optical fibres, using a noise-adapted variant of the SINDy (Sparse Identification of Nonlinear Dynamics) method.
- Automation of physical intuition of non-linear propagation in optical using the Dominant Balance method based on both sparse regression and combinatorical approaches (selected by Optica as a research highlight of 2024).
- Emulation of nonlinear propagation equations using neural networks, particularly an iterative feedforward approach.
2. Experimental results
- Development of a new experimental setup based on sideband truncation and feedback for studying the ideal dynamics of four-wave mixing
- Development of an ultrasensitive real-time spectral measurement method
- Demonstration of active control of supercontinuum generation using a genetic algorithm, both in single-pass fibre and fibre laser setups
- Demonstration of active control of supercontinuum generation for multiphoton microscopy
- Demonstration of a neural network to interpret noise-driven incoherent nonlinear dynamics in optical fibre propagation (selected by Optica as a research highlight of 2025)
Publication summary
The work in OPTIMAL has resulted in a significant publication output:
35 publications in peer-reviewed journals, 30 invited presentations at international peer-reviewed conferences,45 contributions to international peer-reviewed conferences, 15 contributions to national peer-reviewed conferences
A direct link to the HAL database linked to OPTIMAL with access to open-access publications can be found at this link.