Markov Logic Networks with Complex Weights: Expressivity, Liftability and Fourier Transforms

02/24/2020
by   Ondrej Kuzelka, et al.
2

We study expressivity of Markov logic networks (MLNs). We introduce complex MLNs, which use complex-valued weights, and we show that, unlike standard MLNs with real-valued weights, complex MLNs are fully expressive. We then observe that discrete Fourier transform can be computed using weighted first order model counting (WFOMC) with complex weights and use this observation to design an algorithm for computing relational marginal polytopes which needs substantially less calls to a WFOMC oracle than a recent algorithm.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset