Funding Agency

CNPq

Coordinator

Denis Deratani Mauá

Participants

  • Denis Deratani Mauá
  • Fabio G. Cozman, Full Professor at EP-USP
  • Leliane N. de Barros, Associate Professor at IME-USP
  • Thiago P. Bueno, PhD student at IME-USP

Abstract

Bayesian network allow for the compact representation of probabilistic models by exploiting independences among variables. There are however many features that cannot be captured at the variable level; these features are called local structure and include determinism, context-sensitive independence and replication. Inference in Bayesian networks can be made more efficient by exploiting local structure; to this end, we need a convenient formal language for describing (certain types of) local structure. Moreover, we need a language that allows us to speed up inference by combining logical and probabilistic reasoning. This project aims at investigate different formalisms for the specification of Bayesian networks with local structure. We focus on relational languages, particularly, on temporal languages. Our goal is to find languages that are well-suited for describing dynamic features and decrease the complexity of inference.

Publications

  • The complexity of Bayesian networks specified by propositional and relational languages, F. G. Cozman and D. D. Mauá. In Artificial Intelligence, vol. 262 pp. 96–141, 2018.
  • On the Semantics and Complexity of Probabilistic Logic Programs, F. G. Cozman and D. D. Mauá. In Journal of Artificial intelligence Research, vol. 60 pp. 221–262, 2017.
  • Modeling Markov Decision Processes with Imprecise Probabilities Using Probabilistic Logic Programming, T. P. Bueno, D. D. Mauá, L. N. de Barros, and F. G. Cozman. In Proceedings of the Tenth International Symposium on Imprecise Probability: Theories and Applications, vol. 62, pp. 49–60, 2017.
  • The effect of combination functions on the complexity of relational Bayesian networks, D. D. Mauá and F. G. Cozman. In International Journal of Approximate Reasoning, vol. 85 pp. 178–195, 2017.
  • On the complexity of propositional and relational credal networks, F. G. Cozman and D. D. Mauá. In International Journal of Approximate Reasoning, vol. 83 pp. 298–319, 2017.
  • The Descriptive Complexity of Bayesian Network Specifications, F. G. Cozman and D. D. Mauá. In Proceedings of the 14th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty, 2017.
  • The Complexity of Inferences and Explanations in Probabilistic Logic Programming, F. G. Cozman and D. D. Mauá. In Proceedings of the 14th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty, pp. 449–458, 2017.
  • Closed-Form Solutions in Learning Probabilistic Logic Programs by Exact Score Maximization, F. H. O. V. de Faria, F. G. Cozman, and D. D. Mauá. In Proceedings of the 11th International Conference on Scalable Uncertainty Management, pp. 119–133, 2017.