Probabilistic Graphical Models 2: Inference

Probabilistic graphical models are an excellent framework for encoding probability distributions across complex domains. They can be used to encode joint (multivariate), distributions over large numbers random variables that interact with one another. These representations are at the intersection between statistics and computer science. They rely on concepts from probability theory and graph algorithms, machine-learning, and many more. These representations are essential for the development of state-of the-art methods in many applications such as medical diagnosis and image understanding, speech recognition, natural languages processing, and many others. They can also be used to solve many machine learning problems.