@algebraic
@algebraic @kslote1 ·
Got dynamical systems or chaotic PDE data and need governing equations? KANDy: a Kolmogorov-Arnold Network for dynamical system discovery learns interpretable governing equations directly from the most challenging data. 🆕preprint #MachineLearning #DynamicalSystems #SciML #KANB
1
22
ume xelphia
ume xelphia @UmeXelphia ·
総量保存(合計2)の円環モデルで、三点観測が独立でなくなる「三点制約」と、観測空間での閉曲線としての可視化をまとめました。形状保持が破れると軌道が“太る”=診断量になります(玩具モデル)。 DOI: doi.org/10.5281/zenodo… #preprint #dynamicalsystems #math #physics
1
31
F.Mackenzie 约克.小汽车. 嘟嘟
F.Mackenzie 约克.小汽车. 嘟嘟 @FMackenzie7 ·
IC-PINN uses physics-informed neural nets for unsupervised discovery of coupling functions in noisy oscillatory dynamical systems — no direct labels needed, just data + governing equations! #SciML #DynamicalSystems
Jae Kyoung Kim Jae Kyoung Kim @umichkim ·
How do coupled oscillators interact — heartbeats, neurons, circadian clocks? We built IC-PINN to infer coupling functions from data, no basis function selection needed, robust to noise & sparse data doi.org/10.1103/t8g2-n… #DynamicalSystems #ML
1
90