Soliton Solution
Soliton solutions represent stable, self-reinforcing waves that maintain their shape and speed during propagation, a phenomenon studied across diverse fields. Current research focuses on leveraging deep learning, particularly physics-informed neural networks (PINNs) and Fourier neural operators (FNOs), to model and discover soliton dynamics in various nonlinear systems, including the Schrödinger and Korteweg-de Vries equations. This work addresses both forward problems (predicting soliton behavior) and inverse problems (identifying system parameters from observed solutions). These advancements have implications for understanding complex wave phenomena in diverse areas like nonlinear optics, fluid dynamics, and even social dynamics (e.g., modeling information spread).
Papers
Deep learning soliton dynamics and complex potentials recognition for 1D and 2D PT-symmetric saturable nonlinear Schr\"odinger equations
Jin Song, Zhenya Yan
Data-driven localized waves and parameter discovery in the massive Thirring model via extended physics-informed neural networks with interface zones
Junchao Chen, Jin Song, Zijian Zhou, Zhenya Yan