ECE 5412

Bayesian Estimation and Stochastic Optimization

Course Description

Course information provided by the 2019-2020 Catalog.

Covers essential topics in high dimensional statistical inference, stochastic optimization, Bayesian statistical signal processing and Markov Chain Monte-Carlo stochastic simulation. The course is four inter-related parts. Part 1 covers the basics of probabilistic models, Markov chain Monte-Carlo simulation and regression with sparsity constraints. Part 2 covers Bayesian filtering including the Kalman filter, Hidden Markov Model filter and sequential Markov chain Monte-Carlo methods such as the particle filter. Part 3 covers maximum likelihood estimation and numerical methods such as the Expectation Maximization algorithm. Part 4 covers stochastic gradient algorithms  and stochastic optimization. The course focuses on the deep fundamental ideas that underpin signal processing, data science and machine learning - the assignments and project will explore applications.

Outcomes

• Students will learn state of the art methods in Bayesian state estimation, parameter estimation and applications.

When Offered Spring.

View Enrollment Information