On the Role of Fixed Points of Dynamical Systems in Training Physics-Informed Neural Networks.
Franz M. RohrhoferStefan PoschClemens GößnitzerBernhard C. GeigerPublished in: Trans. Mach. Learn. Res. (2023)
Keyphrases
- dynamical systems
- fixed point
- neural network
- state space
- differential equations
- nonlinear dynamical systems
- phase space
- initial conditions
- temporal difference learning
- partially observable
- predictive state representations
- markov chain
- training set
- self organizing maps
- dynamic programming
- support vector
- stationary points
- linear logic
- image sequences
- machine learning