Learning, Logic, and Probability: A Unified View

Abstract

AI systems must be able to learn, reason logically, and handle uncertainty. While much research has focused on each of these goals individually, only recently have we begun to attempt to achieve all three at once. In this talk, I describe Markov logic, a representation that combines the full power of first-order logic and probabilistic graphical models, and algorithms for learning and inference in it. Syntactically, Markov logic is first-order logic augmented with a weight for each formula. Semantically, a set of Markov logic formulas represents a probability distribution over possible worlds, in the form of a Markov network with one feature per grounding of a formula in the set, with the corresponding weight. Formulas and weights are learned from relational databases using inductive logic programming and iterative optimization of a pseudo-likelihood measure. Inference is performed by Markov chain Monte Carlo over the minimal subset of the ground network required for answering the query. Experiments in a real-world university domain illustrate the promise of this approach.