Intelligent near-to-mid-infrared ultrafast fiber lasers with particle swarm-optimized graphene electro-optic modulators

Ultrafast fiber lasers, as core light sources for precision manufacturing, biomedicine, and quantum communication, have long been constrained in performance optimization by complex nonlinear dynamics, environmental sensitivity, and insufficient tuning flexibility. The integration of artificial intelligence (AI) and cutting-edge electro-optic modulation materials offers a transformative paradigm to overcome these bottlenecks. By leveraging the Fermi level tuning property of graphene and embedding the particle swarm optimization (PSO) algorithm, we developed an intelligent electro-optic modulation device that enables closed-loop dynamic control of ultrafast fiber lasers at 1.5 and 2.8 μm. Through closed-loop optimization by the particle swarm algorithm, rapid self-starting and termination of pulse states, as well as programmable and reversible control of pulse width, repetition rate, and output power were achieved. In addition, our system can automatically identify and resolve pulse splitting caused by external disturbances. We establish broadband intelligent electro-absorption in the near-to-mid infrared spectral range, laying the technical foundation for broadband infrared photonics in the 1–3 μm range.