Bayesian Networks 2 – Forward-Backward – Stanford CS221: AI

Stanford University via YouTube

YouTube based on 0 ratings

Likes

Overview

Bayesian Networks 2 - Forward-Backward is an advanced module from Stanford’s CS221 Artificial Intelligence course that delves into the forward-backward algorithm, a key technique for performing efficient inference in dynamic Bayesian networks such as Hidden Markov Models (HMMs). This course builds on foundational knowledge of Bayesian networks and inference, focusing on temporal models that handle sequences of data over time. You will learn how the forward-backward algorithm computes the probabilities of hidden states at each time step by combining information from...

Syllabus

Introduction. Review: Bayesian network. Review: probabilistic inference. Hidden Markov model inference. Lattice representation. Summary. Hidden Markov models. Review: beam search. Step 1: propose. weight. Step 3: resample. Application: object tracking. Particle filtering demo. Roadmap. Gibbs sampling.