You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
List of papers and content relevant to ODINN. You can comment here to make new suggestions and I will add them to this first post to make the list grow.
This is the baseline paper for all this work. This describes the method to combine mechanistic models in the form of differential equations with neural networks, in order to learn subparts an equation. This opens the door to learning missing information in physical models, and to discover new laws and parametrizations.
This paper analyses the difficulties in training neural differential equations (NDEs) with stiff differential equations. The basic approach from the Neural ODEs paper using a Backsolve adjoint doesn't work except for the simplest cases. In order to correctly obtain gradients for more complex systems, Interpolating or Quadrature adjoints are presented, which are more stable and robust. Other strategies are discussed on how to successfully train these harder NDEs.
Benchmark of the different backpropagation techniques for Universal Differential Equations. Goes in the same direction as the "Stiff Neural Ordinary Differential Equations" paper above.
Paper describing SINDy, the method used to derive a mathematical expression from nonlinear systems (in our case, neural networks). This will be used as the last step in order to transform the learnt nonlinear functions by neural networks into mathematical expressions which can be used to rewrite the original equation of the physical process.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
List of papers and content relevant to ODINN. You can comment here to make new suggestions and I will add them to this first post to make the list grow.
Beta Was this translation helpful? Give feedback.
All reactions