Abstract

Markov Logic Networks (MLNs) are well-suited for expressing statistics suchas "with high probability a smoker knows another smoker" but not for expressingstatements such as "there is a smoker who knows most other smokers", which isnecessary for modeling, e.g. influencers in social networks. To overcome thisshortcoming, we study quantified MLNs which generalize MLNs by introducingstatistical universal quantifiers, allowing to express also the latter type ofstatistics in a principled way. Our main technical contribution is to show thatthe standard reasoning tasks in quantified MLNs, maximum a posteriori andmarginal inference, can be reduced to their respective MLN counterparts inpolynomial time.

 

Quick Read (beta)