Cubature Kalman Filter Based Training of Hybrid Differential Equation Recurrent Neural Network Physiological Dynamic Models.
Ahmet DemirkayaTales ImbiribaKyle LockwoodSumientra RampersadElie AlhajjarGiovanna GuidoboniZachary DanzigerDeniz ErdogmusPublished in: EMBC (2021)
Keyphrases
- dynamic model
- recurrent neural networks
- differential equations
- recurrent networks
- echo state networks
- feedforward neural networks
- neural network
- experimental data
- dynamical systems
- neural network structure
- boundary value problem
- control scheme
- hidden layer
- initial conditions
- numerical solution
- feed forward
- reservoir computing
- neural model
- complex valued
- numerical methods
- training algorithm
- multiple models
- ordinary differential equations
- training process
- artificial neural networks
- training phase
- continuous functions
- dynamic programming
- numerical integration
- nonlinear differential equations
- biologically inspired
- machine learning