Inverse prediction of ultrashort pulse nonlinear propagation in optical fibers

Recently, artificial intelligence has been proven as an effective modeling tool in ultrafast optics; its application in the design of ultrafast laser systems is a promising issue. In this Letter, a method based on a feed-forward neural network(FNN) model with a simple structure is adopted to inversely predict the full-field supercontinuum generation and recover the initial pulse.The performance of the FNN and its dependence on the predicted pulse features are further explored by a reconstruction test. The generalization ability of the proposed method is further demonstrated in the case with an initial chirp.