Samuel Kincel

Activation Function Effects in Physics-Informed Neural Networks for One-Dimensional Signal Prediction

Physics-informed neural networks provide a mesh-free way of solving differential equations by embedding governing physical laws directly into the training objective. This paper evaluates their suitability for one-dimensional signal prediction governed by representative partial differential equations. Three equations wi...