2026-10-06T05:10:04Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34046484 2026-10-01T15:04:42Z category_29164 category_28813 item_type_1 month_year_10_2026

Hyperparameter optimization results for the PIKAN model Gregorio Perez Bernal (22258905) Oscar Rincón Cardeño (22258780) Silvana Montoya-Noguera (22258901) Nicolas Guarin-Zapata (6532391) Neural networks Applications in physical sciences Scientific machine learning (SciML) Physics-informed neural networks (PINNs) Kolmogorov-Arnold Networks(KAN) Inverse problems for differential equations Unbounded Domains (UDs) <p dir="ltr">The top panel shows the mean global error obtained in each trial, with the solid line indicating the best objective value achieved up to each trial. The lower panels show the corresponding error as a function of the number of hidden layers L, neurons per layer N, grid size G, polynomial order p, and learning rate $\alpha$. The color of each marker indicates the trial number, illustrating the progression of the optimization toward lower-error configurations. All error values are shown on a logarithmic scale</p> 2026-10-01T15:04:42Z Image Figure 10.6084/m9.figshare.34046484.v1 https://figshare.com/articles/figure/Hyperparameter_optimization_results_for_the_PIKAN_model/34046484 CC BY 4.0