Welcome to SciML’s documentation!¶
It is a repository to experiment Scientific Machine Learning (SciML) in simulating physical dynamics, understanding machine learning pros and cons in scientific computing, and discovering physical rules using the data-driven and physics-based method.
The fundamental crux of the project is to solve a variety of differential equations with machine learning.
We studied the following physical phenomenons:
Pendulum
Spring Mass
Wave Propagation
Poisson
Lorenz
with the following SciML models:
Physics Informed NN PINN¶
Link: https://nips.cc/
Neural ODE (NODE)¶
Univeral Differential Equations (UDE)¶
Hamiltonian Neural Network (HNN)¶
Hamiltonian fundamentals:
Link: http://www.scholarpedia.org/article/Hamiltonian_systems