Generative structure learning for Markov Logic Networks.

Abstract : In this paper, we present a generative algorithm to learn Markov Logic Network (MLN) structures automatically, directly from a training dataset. The algorithm follows a bottom-up approach by first heuristically transforming the training dataset into boolean tables, then creating candidate clauses using these boolean tables and finally choosing the best clauses to build the MLN. Comparisons to the state-of-the-art structure learning algorithms for MLNs in two real-world domains show that the proposed algorithm outperforms them in terms of the conditional log likelihood (CLL), and the area under the precision-recall curve (AUC).
Document type :
Conference papers
Liste complète des métadonnées
Contributor : Christel Vrain <>
Submitted on : Monday, July 19, 2010 - 5:26:43 PM
Last modification on : Thursday, January 17, 2019 - 3:06:04 PM


  • HAL Id : hal-00504074, version 1



Quang-Thang Dinh, Matthieu Exbrayat, Christel Vrain. Generative structure learning for Markov Logic Networks.. STAIRS 2010, fifth European Starting AI Researcher Symposium., Aug 2010, Lisbon, Portugal. pp.63-75. ⟨hal-00504074⟩



Record views