Discriminative Probabilistic Models for Relational Data

Outline

nBayesian Networks

nProbabilistic Relational Models

nCollective Classification & Clustering

nUndirected Discriminative Models

nCollective Classification Revisited

nPRMs for NLP

nWord-Sense Disambiguation

nRelation Extraction

nNatural Language Understanding (?)

* An outsider’s perspective

or “Why Should I Care?”*


	We will look at 2 classes of prob graphical models that represent complex distributions compactly and allow us to reason
	efficiently. First, we will look at directed models, called Bayesian networks and their extension to relational data, PRMs. We’ll then consider
	undirected models, such as Markov nets and Relational Markov nets.