State Space Models: A Modern Approach
State Space Models: A Modern Approach
Scratchpad
Introduction
Hidden Markov Models
Hidden Semi-Markov Models
Linear Gaussian SSMs
Non-Linear Gaussian SSMs
Non-Gaussian SSMs
Switching SSMs
Deep SSMs
Recurrent Neural Networks
Inference in discrete SSMs
HMM filtering (forwards algorithm)
HMM smoothing (forwards-backwards algorithm)
Viterbi algorithm
Parallel HMM smoothing
Forwards-filtering backwards-sampling algorithm
Inference in linear-Gaussian SSMs
Kalman filtering
Kalman (RTS) smoother
Parallel Kalman Smoother
Forwards-filtering backwards sampling
Linearization methods for non-linear Gaussian SSMs
Extended Kalman filtering
Extended Kalman smoother
Parallel extended Kalman smoothing
Linearization methods for non-linear Gaussian SSMs
Unscented filtering
Unscented smoothing
Quadrature and cubature methods
Posterior linearization
Assumed Density Filtering
Variational inference
Particle filtering
Sequential Monte Carlo
Offline parameter estimation (learning)
Expectation Maximization (EM)
Stochastic Gradient Descent (SGD)
Variational Bayes (VB)
Markov Chain Monte Carlo (MCMC)
Multi-target tracking
Data assimilation using Ensemble Kalman filter
Changepoint detection
Timeseries forecasting
Markovian Gaussian processes
Differential equations and SSMs
Optimal control
Bibliography
.ipynb
.pdf
repository
open issue
Binder
Colab
Expectation Maximization (EM)
Expectation Maximization (EM)
ΒΆ
[
GH96
]
previous
Offline parameter estimation (learning)
next
Stochastic Gradient Descent (SGD)