APEBench是由慕尼黑工业大学的机器学习中心创建的一个综合性基准套件，旨在评估自回归神经网络在求解偏微分方程（PDE）中的表现。该数据集包含46种不同的PDE动态，涵盖从一维到三维的空间维度，数据量庞大且多样化。数据集的创建过程结合了高效的伪谱方法和可微分模拟框架，确保了数据生成的高效性和准确性。APEBench主要应用于神经网络架构的评估和优化，旨在解决复杂PDE模拟中的精度和速度问题。

arXiv2024-11-01 更新4260

Data from: Multifractal evidence of nonlinear interactions stabilizing posture for phasmids in windy conditions: a reanalysis of insect postural-sway data

生物学

非线性动力学

The present work is a reanalysis of prior work documenting postural sway in phasmids (i.e., “stick insects”) [1]. The prior work pursued the possibility that postural sway was an evolutionary adaptati

DataONE2018-08-24 更新150

Dynamical mechanism of Lévy flight driven by the nonlinear friction

莱维飞行

非线性动力学

Dynamical mechanism of Lévy flight driven by the nonlinear friction

Digital Science - Figshare2026-03-27 更新80

PDE dataset

偏微分方程

机器学习

Operator learning involving parametric partial differential equations

kaggle2021-06-27 更新570

Appendix A. Derivation of the partial differential equation (PDE) for U(w,t).