Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Abril 2023
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Temario
1 Motivación
2 Problema a modelar
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Motivación
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
z0 =t
z1 = σ(W1 z0 + b1 )
z2 = σ(W2 z1 + b2 )
z3 = W3 z2 + b3
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
r = ẍ + 2ζω0 ẋ + ω02 x = 0
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
r = ẍ + 2ζω0 ẋ + ω02 x = 0
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Situación a modelar
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento via redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento via redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento via redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a redes neuronales
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento via inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Usando el mismo espacio parametral del modelo anterior
Θ = {θ, ϑ, ϵ}, estimamos P(Θ), mediante:
P(Θ) = P(c, k, x0 , Wk , bk )
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Usando el mismo espacio parametral del modelo anterior
Θ = {θ, ϑ, ϵ}, estimamos P(Θ), mediante:
P(Θ) = P(c, k, x0 , Wk , bk )
Donde
log (c) ∼ N (µ = log (2.2), σ = 0.5)
log (k) ∼ N (µ = log (350), σ = 0.5)
log (x0 ) ∼ N (µ = log (0.56), σ = 0.5)
Wk ∼ N (µ = 0.0, σ = 2.0)
bk ∼ N (µ = 0.0, σ = 2.0)
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Modelamiento vı́a inferencia Bayesiana
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Conclusiones del artı́culo
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Conclusiones del artı́culo
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Conclusiones del artı́culo
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Conclusiones del artı́culo
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst
Gracias
Juan David Carrascal Ibañez Bayesian Physics Informed Neural Networks for real-world nonlinear dynamical syst