Deep reinforcement learning algorithm for self-tuning 8-figure fiber laser

Machine learning (ML) algorithms have already shown their efficiency for adjusting fiber-mode-locked lasers [1] . However, the performance of reinforcement (RL) learning algorithms required for robust application of ML methods in practical environment is yet to be verified in different laser systems [2] . Implementation of RL algoritms may reveal unknown strategies for adjusting complex laser systems since such algorithms consider intermidiate states of the system during the training process.