AI-enabled Fast and Accurate Modeling for Femtosecond Chirped-pulse Amplification

Modeling the propagation of femtosecond pulses in the gain fiber of a chirped-pulse amplification (CPA) system using the traditional method is rather computationally intensive. Here, we propose a long short-term memory (LSTM) model to accelerate the pulse propagation in the gain fiber of an emulated CPA system. The LSTM model can accurately simulate the pulse propagation in the gain fiber and is ~84 times faster than the traditional method under an identical hardware platform.