CI-PINN: Causal Integral Physics-Informed Neural Network for Solving Evolution Equations

License: See LICENSE file · Updated 2026-09-29
My rating

The resource presents a research paper on physics-informed neural networks for solving evolution equations and partial differential equations. It introduces CinNet, whose Volterra-type causal integral module explicitly represents temporal dependence that conventional pointwise space-time PINNs do not encode directly. The method is designed to improve prediction accuracy for systems governed by evolution equations.

Tags: