Room 408, College of Science, NUK
(高雄大學理學院 408室)
Numerical Simulations and Parameter Identification via Statistical and Neural Network Methods
I-Yun Cheng (National University of Kaohsiung)
Abstract:
In this talk, we present extensive numerical simulations to validate the theoretical predictions obtained from amplitude equations. The results confirm the coexistence and transitions of hexagonal and stripe patterns under varying diffusion and control parameters. Furthermore, we introduce two distinct approaches for parameter identification based on Turing patterns: a statistical inference method using empirical cumulative distribution functions and a deep learning approach employing a convolutional neural network (CNN) inspired by the VGG16 framework. The CNN-based method demonstrates higher accuracy and efficiency, offering a powerful tool for linking theoretical models to empirical spatial data.
Organizers:Chih-Hung Chang (NUK) & Yu-Hau Liang (NUK)
