Abstract:Deep neural networks (DNNs) have achieved exceptional performance across various fields by learning complex nonlinear mappings from large-scale datasets. However, they encounter challenges such as high computational costs and limited interpretability. To address these issues, hybrid approaches that integrate physics with AI are gaining interest. This paper introduces a novel physics-based AI model called the "Nonlinear Schrödinger Network", which treats the Nonlinear Schrödinger Equation (NLSE) as a general-purpose trainable model for learning complex patterns including nonlinear mappings and memory effects from data. Existing physics-informed machine learning methods use neural networks to approximate the solutions of partial differential equations (PDEs). In contrast, our approach directly treats the PDE as a trainable model to obtain general nonlinear mappings that would otherwise require neural networks. As a type of physics-AI symbiosis, it offers a more interpretable and parameter-efficient alternative to traditional black-box neural networks, achieving comparable or better accuracy in some time series classification tasks while significantly reducing the number of required parameters. Notably, the trained Nonlinear Schrödinger Network is interpretable, with all parameters having physical meanings as properties of a virtual physical system that transforms the data to a more separable space. This interpretability allows for insight into the underlying dynamics of the data transformation process. Applications to time series forecasting have also been explored. While our current implementation utilizes the NLSE, the proposed method of using physics equations as trainable models to learn nonlinear mappings from data is not limited to the NLSE and may be extended to other master equations of physics.

Neural Network Emulation of Spatio-temporal Data Using Linear and Nonlinear Dimensionality Reduction

A survey of data-driven emulators for high dimensional spatio-temporal data

Manifold Learning for the Emulation of Spatial Fields from Computational Models.

Reduced Dimensional Gaussian Process Emulators Of Parametrized Partial Differential Equations Based On Isomap

A Decentralized Approach for Nonlinear Prediction of Time Series Data in Sensor Networks

Nonlinear Schrödinger Network

Temporal Subsampling Diminishes Small Spatial Scales in Recurrent Neural Network Emulators of Geophysical Turbulence

Less is More: a New Machine Learning Methodology for Spatiotemporal Systems

Learning the intrinsic dynamics of spatio-temporal processes through Latent Dynamics Networks

Neural-network learning of SPOD latent dynamics

Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning

A Note on Dimensionality Reduction in Deep Neural Networks using Empirical Interpolation Method

Data-driven reduced-order modeling of spatiotemporal chaos with neural ordinary differential equations

The Neural Network shifted-proper orthogonal decomposition: A machine learning approach for non-linear reduction of hyperbolic equations

Equivariant Neural Simulators for Stochastic Spatiotemporal Dynamics

Learning nonlinear operators in latent spaces for real-time predictions of complex dynamics in physical systems

Nonparametric Estimation for High-Dimensional Space Models Based on a Deep Neural Network

Data-driven spatiotemporal modeling for structural dynamics on irregular domains by stochastic dependency neural estimation

Nonlinear dimensionality reduction using a temporal coherence principle

Non-intrusive model reduction of advection-dominated hyperbolic problems using neural network shift augmented manifold transformations

Dense ReLU Neural Networks for Temporal-spatial Model