Abduction is inference to the best explanation of a given set of evidence. It is important for plan or intent recognition systems. Traditional approaches to abductive reasoning have either used first-order logic, which is unable to reason under uncertainty, or Bayesian networks, which can handle uncertainty using probabilities but cannot directly handle an unbounded number of related entities. This paper proposes a new method for probabilistic abductive reasoning that combines the capabilities of first-order logic and graphical models by using Markov logic networks. Experimental results on a plan recognition task demonstrate the effectiveness of this method.

PDF

In Proceedings of the IJCAI-09 Workshop on Plan, Activity, and Intent Recognition (PAIR-09), Pasadena, CA, July 2009.

@inproceedings{kate:pair09, title={Probabilistic Abduction using Markov Logic Networks}, author={Rohit J. Kate and Raymond J. Mooney}, booktitle={Proceedings of the IJCAI-09 Workshop on Plan, Activity, and Intent Recognition (PAIR-09)}, month={July}, address={Pasadena, CA}, url="http://www.cs.utexas.edu/users/ai-lab/?kate:pair09", year={2009} }

Slides