Swift Markov Logic for Probabilistic Reasoning on Knowledge Graphs

This repository contains the full report of the experimental results (execution time and error rates) and the datasets used to evaluate the probabilistic reasoning algorithm—MCMC-Chase—of Soft Vadalog.

Soft Vadalog is the probabilistic extension of the Vadalog System, a state-of-the-art reasoner for Knowledge Graphs. A detailed discussion of the system architecture can be found here.

A preliminary presentation of MCMC-Chase has been recently published at RuleML+RR 2020.

Execution Times and Error Rates

Access via Google Spreadsheets

Datasets

1. BASE nodes=100
2. BASE nodes=250
3. BASE nodes=500
4. BASE nodes=1000
5. DENSE nodes=100
6. DENSE nodes=250
7. DENSE_nodes=500
8. DENSE_nodes=1000
9. SUPERDENSE nodes=100
10. SUPERDENSE nodes=250
11. SUPERDENSE nodes=500
12. SUPERDENSE nodes=1000

Authors and Contacts

• Luigi Bellomarini (Banca d’Italia)
• Eleonora Laurenza (Banca d’Italia)
• Emanuel Sallinger (TU Wien & University of Oxford)
• Evgeny Sherkhonov (University of Oxford)