Parameter Space Compression and Random Structure Automatic Generation for the inverse design of Photonic Crystal Fibers Based on Convolutional Adversarial Autoencoder

An automatic iterative optimization and inverse design method for photonic crystal fiber (PCF) structures is proposed based on convolutional adversarial autoencoder (CAAE) and forward prediction neural network. This method takes the two-dimensional (2D) material refractive index distribution of the fiber cross-section as the structural parameter space and can achieve automatic generation and iterative optimization of PCF structures in the 2D parameter space. The automatic generation of random PCF structures is based on CAAE network, which can compress the 2D fiber structural parameter matrix into a 36-dimensional hyperparameter space with Gaussian distribution, and the Gaussian hyperparameters can be restored to the original 2D structural parameter matrix through the decoder. The decoder can work independently and generate random PCF structures after inputting random Gaussian hyperparameters. Then, a forward prediction convolutional neural network (PCNN) is constructed to evaluate the optical property of the PCFs generated from the decoder, including mode effective refractive index, dispersion and mode field area. The input variable of this PCNN is the 2D fiber structure matrix. By combining the PCNN and CAAE networks, automatic generation and iterative optimization of 2D PCF structures can be achieved. The structural variables generated and optimized are the 2D refractive index distribution of the fiber cross-section, not fixed specific structural parameters, which is more flexible and can be applied to different types of PCFs. We also propose a transfer learning method for the random generation of different PCF structures based on CAAE network, which only needs a small amount of data to train the autoencoder, and efficient and accurate random generation of other PCFs with different lattice arrangements can be achieved. The proposed automatic generation and iterative optimization method is flexible and efficient, which provide a new approach for the optimization and reverse design of PCF structures in 2D parameter space.