Machine Learning Enabled Directly Modulated Lasers
Résumé
The complex dynamics of systems based on directly modulated lasers (DMLs) can hinder their optimization, as their simulation entails numerical methods where gradient calculation is not readily available. Our proposed solution combines data-driven DML modeling with end-to-end (E2E) learning techniques to obtain optimal link configurations.