Title: Learning Markov Logic Network Structure by Template Constructing
Date/Time: April 14th, 2017 @ 3:00 PM
Place: 223 Atanasoff
Major Professor: Jin Tian
Committee Members: Pavankumar Aduri, Carl Chang
Abstract:
Markov logic networks (MLNs) are a statistical relational model that incorporates first-order logic and probability by attaching weights to first-order clauses. However, due to the large search space, the structure learning of MLNs is a computationally expensive problem. In this paper, we present a new algorithm for learning the structure of Markov Logic Network by directly utilizing the data to construct the candidates. Our approach makes use of a Markov Network learning algorithm to construct a “template” network. We then apply the template to guide the candidate clauses construction process. The results of experiment demonstrate that our algorithm is promising.