A Gaussian Approximation for Stochastic Nonlinear Dynamical Processes with Annihilation

Posted in Science on August 22, 2008


A Gaussian Approximation for Stochastic Nonlinear Dynamical Processes with Annihilation

Lecture slides:

  • A Gaussian approximation for stochastic non-linear<br> dynamical processes with annihilation
  • Contents
  • Stochastic optimal control problem
  • Hamilton-Jacobi-Bellman equation
  • Log transformation and optimal control
  • Fokker-Planck with decay
  • A stochastic dynamical process with annihilation
  • Relation with discrete time Kalman smoothing
  • The transition density
  • Path integral formulation
  • Euler-Lagrange equations
  • Euler-Lagrange equations01
  • A formal forward-backward algorithm
  • Numerical example
  • Relation with classical mechanics (b = 0)
  • Size of fluctuations: linear noise approximation
  • Numerical example
  • Partition function
  • Partition function: numerical result
  • Summary
  • Discussion

Author: Wim Wiegerinck, Radboud University Nijmegen

Watch Video

Tags: Science, Lectures, Computer Science, Machine Learning, VideoLectures.Net, Gaussian Processes