Extending Markov Logic to Model Probability Distributions in Relational Domains

Authors:
Dominik Jain;Bernhard Kirchlechner;Michael Beetz
Affiliations:
Intelligent Autonomous Systems Group, Department of Informatics, Technische Universität München,;Intelligent Autonomous Systems Group, Department of Informatics, Technische Universität München,;Intelligent Autonomous Systems Group, Department of Informatics, Technische Universität München,
Venue:
KI '07 Proceedings of the 30th annual German conference on Advances in Artificial Intelligence
Year:
2007

Citing 8
Cited 5

Probabilistic logic

Artificial Intelligence
Representing and reasoning with probabilistic knowledge: a logical approach to probabilities

Representing and reasoning with probabilistic knowledge: a logical approach to probabilities
Learning Probabilistic Relational Models

IJCAI '99 Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence
Dependency Networks for Relational Data

ICDM '04 Proceedings of the Fourth IEEE International Conference on Data Mining
Markov logic networks

Machine Learning
An analysis of first-order logics of probability

IJCAI'89 Proceedings of the 11th international joint conference on Artificial intelligence - Volume 2
Visually tracking football games based on TV broadcasts

IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
BLOG: probabilistic models with unknown objects

IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence

Towards performing everyday manipulation activities

Robotics and Autonomous Systems
Adaptive Markov Logic Networks: Learning Statistical Relational Models with Dynamic Parameters

Proceedings of the 2010 conference on ECAI 2010: 19th European Conference on Artificial Intelligence
Soft evidential update via Markov chain Monte Carlo inference

KI'10 Proceedings of the 33rd annual German conference on Advances in artificial intelligence
A hybrid deliberative layer for robotic agents: fusing DL reasoning with HTN planning in autonomous robots

A hybrid deliberative layer for robotic agents: fusing DL reasoning with HTN planning in autonomous robots
Bayesian logic networks and the search for samples with backward simulation and abstract constraint learning

KI'11 Proceedings of the 34th Annual German conference on Advances in artificial intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

Markov logic, as a highly expressive representation formalism that essentially combines the semantics of probabilistic graphical models with the full power of first-order logic, is one of the most intriguing representations in the field of probabilistic logical modelling. However, as we will show, models in Markov logic often fail to generalize because the parameters they contain are highly domain-specific. We take the perspective of generative stochastic processes in order to describe probability distributions in relational domains and illustrate the problem in this context by means of simple examples.We propose an extension of the language that involves the specification of a priori independent attributes and that furthermore introduces a dynamic parameter adjustment whenever a model in Markov logic is instantiated for a certain domain (set of objects). Our extension removes the corresponding restrictions on processes for which models can be learned using standard methods and thus enables Markov logic networks to be practically applied to a far greater class of generative stochastic processes.