Please clarify usage of FACT++ reasoner in BUNDLE

[Konrad Abicht]

I am conducting a small study and would like to test BUNDLE reasoner. My findings say that BUNDLE is a stand-alone reasoner to some extent, for instance:

RIGUZZI, Fabrizio, et al. BUNDLE: A reasoner for probabilistic ontologies. In: Web Reasoning and Rule Systems: 7th International Conference, RR 2013, Mannheim, Germany, July 27-29, 2013. Proceedings 7. Springer Berlin Heidelberg, 2013. S. 183-197.
PDF: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=763f3c5f35e6a04be3ef0faed626ccb3168e13b0

There is no mention of FACT++, but when browsing the source code as well as looking into the downloadable files, I find a lot of references about it. Can you tell me, where FACT++ is used in BUNDLE? Is it used for special tasks? Or is it replacing old BUNDLE related tasks?

In CHANGELOG file it is written that BUNDLE supports FACT++ and other reasoners since v3.0. What does support means here?

Thank you for your time in advance.

[Riccardo Zese]

Dear Konrad,

thank you for your email.

From the paper you mention, BUNDLE has been improved a lot. In the last years, we extended its core reasoning system so that BUNDLE can use other reasoners besides Pellet. Basically, BUNDLE encapsulates different reasoners to which it demands the computation of the explanations for the given query. Then, it uses the explanation found to compute the probability of the query in the case of probabilistic KB. Nowadays, you can choose between Pellet, HermiT, JFact and Fact++ as the underlying reasoner to find the explanations. You can also use TRILL, which is written in Prolog, so its installation requires a little more work (you need to install SWI-Prolog, JPL library and add to SWI-Prolog the TRILL pack).

You can choose the reasoner by using the argument

• -r,--reasoner <pellet | hermit | jfact | fact++>

‌

You can find more up-to-date information at https://ml.unife.it/bundle/. In particular, the paper that describes the modularity of BUNDLE is

Giuseppe Cota, Fabrizio Riguzzi, Riccardo Zese, Elena Bellodi, and Evelina Lamma. A Modular Inference System for Probabilistic Description Logics. In Scalable Uncertainty Management. Springer International Publishing, 2018. pages 78-92. [bib | DOI | .pdf]

‌

Hope this clarifies your doubts. Otherwise don’t hesitate to ask again!

‌

Riccardo

‌

[Riccardo Zese]

To answer the question added in the edit of your question, support means that BUNDLE can use that reasoner.

[Konrad Abicht]

Many thanks for your quick response! I will look into that. Issue can be closed for now.

[Riccardo Zese]

Let me know if you need something else.