Neural Ordinary Differential Equations
PulseAugur coverage of Neural Ordinary Differential Equations — every cluster mentioning Neural Ordinary Differential Equations across labs, papers, and developer communities, ranked by signal.
1 天有情绪数据
-
新的NHODE框架学习具有未观测状态的物理信息动力学系统
研究人员开发了一个名为神经哈密顿常微分方程（NHODE）的新框架，用于从数据中学习动力学系统，即使某些状态变量未被观测到。该方法结合了哈密顿神经网络和神经ODE，嵌入了能量守恒等物理结构，以提高泛化能力和稳定性。NHODE框架在包括混沌三体问题在内的各种系统上进行了测试，与纯数据驱动的方法相比，在精度和长期预测能力方面表现更优。
-
Holomorphic KAN-ODE 模型以可解释方程模拟复杂动力学
研究人员开发了一个名为 Holomorphic KAN-ODE 的新框架，将 Kolmogorov-Arnold Networks (KANs) 集成到神经常微分方程 (Neural ODEs) 中。该方法通过纳入复分析先验并遵守 Cauchy-Riemann 条件，旨在更好地模拟具有分形边界的复杂动力学系统。与传统的 MLP 相比，Holomorphic KAN-ODE 框架表现出卓越的性能，在重建动力学系统、识别控制方程以及提高对…
-
Neural ODEs learn quantum many-body dynamics, guiding closure scheme development
Researchers have developed a neural ordinary differential equation (ODE) model capable of simulating the dynamics of quantum many-body systems. This model, trained on exact two-particle reduced density matrix (2RDM) dat…
-
PINN-Cast transformer uses Neural ODEs and physics loss for weather forecasting
Researchers have developed PINN-Cast, a novel continuous-depth transformer model for short-term weather forecasting. This model integrates Neural Ordinary Differential Equations (Neural ODEs) within its encoder blocks t…