Re: p4-feedback Digest, Vol 88, Issue 14

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Re: p4-feedback Digest, Vol 88, Issue 14

Eddy Vanderlinden
Thanks a lot Marco, Timothy,

I'll inform the university of Oxford of the problem and use Pellet and
Fact++ meanwhile.

My small test ontology with such a small number of individuals (27
individuals in 3 classes) should not be a problem to any reasoner I hope.

Best regards

Eddy

-----Oorspronkelijk bericht-----
From: [hidden email]
Sent: Saturday, January 25, 2014 9:01 PM
To: [hidden email]
Subject: p4-feedback Digest, Vol 88, Issue 14

Send p4-feedback mailing list submissions to
[hidden email]

To subscribe or unsubscribe via the World Wide Web, visit
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
or, via email, send a message with subject or body 'help' to
[hidden email]

You can reach the person managing the list at
[hidden email]

When replying, please edit your Subject line so it is more specific
than "Re: Contents of p4-feedback digest..."


Today's Topics:

   1. Re: Problems with Hermit reasoner (Timothy Redmond)
   2. Re: Problems with Hermit reasoner (Marco Colombetti)


----------------------------------------------------------------------

Message: 1
Date: Fri, 24 Jan 2014 17:14:13 -0800
From: Timothy Redmond <[hidden email]>
To: [hidden email]
Subject: Re: [p4-feedback] Problems with Hermit reasoner
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=UTF-8; format=flowed


I am not an expert on reasoners but it is not surprising to me that this
type of ontology gives many current OWL reasoners a lot of problems.  I
believe that, to a first approximation, many owl reasoners essentially
work by building a model of an ontology. There are some issues such as
having a method of finding and encoding models that are infinite.

To the extent that this is a valid explanation of how OWL reasoners
work, I can see why a reasoner may have trouble building these models.
If, during the course of building a model for your ontology, if a
reasoner is looking at at an individual that is a member of a class
expression like

       isAttendedBy min 13 Personnel


the reasoner needs to find or assert 13 different individuals in the
Personnel class.  This growth in the size of the model may continue
exponentially.

Regardless of the approach taken by a reasoner, it is possible to find
small ontologies that make a reasoner think for a very long time.  The
description logic complexity navigator,

         http://www.cs.man.ac.uk/~ezolin/dl/


shows that the worst case performance of inference over many classes of
ontologies may be very bad.  The positive claim is made is that most
ontologies in general use do much better than this worst case scenario.

-Timothy


On 01/23/2014 11:50 PM, Eddy Vanderlinden wrote:

> Dear Prot?g? team,
>
> Could you please let me know if you can help with the problem described
> here
> or wether I should turn to the producers of Hermit reasoner?
>
> In the small test ontology added to the post there is class
> BPMNTrainingFeasible.
> The inferences are working properly with Fact++, not with Hermit.
>
>
> In Prot?g? 4.3, reasoner Hermit :
> The equivalence description:
> BPMNModellingTraining
>   and (isAttendedBy min 6 Personnel)
> Makes the reasoner work for more than 3 hours.
>
> The equivalence description:
> BPMNModellingTraining
>   and (isAttendedBy min 13 Personnel)
> makes the reasoner work for more than 23 hours.
> After that, the reasoner was stopped.
> The explanation could not be obtained.
>
> In Prot?g? 4.3, reasoner Pellet:
> The equivalence description:
> BPMNModellingTraining
>   and (isAttendedBy min 13 Personnel)
> makes the reasoner work for not more than 147 ms.
> The explanatioins were more than 200 and stopped at that time.
>
> In Prot?g? 4.1, reasoner Pellet:
> The equivalence description:
> BPMNModellingTraining
>   and (isAttendedBy min 6 Personnel)
> Makes the reasoner work for not more than 18 ms.
>
> The equivalence description:
> BPMNModellingTraining
>   and (isAttendedBy min 13 Personnel)
> makes the reasoner work for not more than 167 ms.
> However the reasoner had to be stopped and restarted to get the correct
> result.
>
> The explanation for the inferred members could be obtained in 86 ms.
>
> test.owl <http://protege-project.136.n4.nabble.com/file/n4659770/test.owl>
>
> Thanks in advance for any direction.
>
> Best regards
>
> Eddy
>
>
>
>
> --
> View this message in context:
> http://protege-project.136.n4.nabble.com/Problems-with-Hermit-reasoner-tp4659770.html
> Sent from the Protege-OWL 4.x Support mailing list archive at Nabble.com.
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback



------------------------------

Message: 2
Date: Sat, 25 Jan 2014 18:56:35 +0000
From: Marco Colombetti <[hidden email]>
To: Protege 4.x support and discussion
<[hidden email]>
Subject: Re: [p4-feedback] Problems with Hermit reasoner
Message-ID: <CF09C71A.8516%[hidden email]>
Content-Type: text/plain; charset="iso-8859-1"

I also had problems with HermiT with cardinality restrictions involving
numbers greater than 3 or 4.

I was explained it depends on how cardinality restrictions are
represented, which is different from Pellet and Fact++.

Marco



On 24.01.2014 08:50 , "Eddy Vanderlinden" <[hidden email]>
wrote:

>Dear Prot?g? team,
>
>Could you please let me know if you can help with the problem described
>here
>or wether I should turn to the producers of Hermit reasoner?
>
>In the small test ontology added to the post there is class
>BPMNTrainingFeasible.
>The inferences are working properly with Fact++, not with Hermit.
>
>
>In Prot?g? 4.3, reasoner Hermit :
>The equivalence description:
>BPMNModellingTraining
> and (isAttendedBy min 6 Personnel)
>Makes the reasoner work for more than 3 hours.
>
>The equivalence description:
>BPMNModellingTraining
> and (isAttendedBy min 13 Personnel)
>makes the reasoner work for more than 23 hours.
>After that, the reasoner was stopped.
>The explanation could not be obtained.
>
>In Prot?g? 4.3, reasoner Pellet:
>The equivalence description:
>BPMNModellingTraining
> and (isAttendedBy min 13 Personnel)
>makes the reasoner work for not more than 147 ms.
>The explanatioins were more than 200 and stopped at that time.
>
>In Prot?g? 4.1, reasoner Pellet:
>The equivalence description:
>BPMNModellingTraining
> and (isAttendedBy min 6 Personnel)
>Makes the reasoner work for not more than 18 ms.
>
>The equivalence description:
>BPMNModellingTraining
> and (isAttendedBy min 13 Personnel)
>makes the reasoner work for not more than 167 ms.
>However the reasoner had to be stopped and restarted to get the correct
>result.
>
>The explanation for the inferred members could be obtained in 86 ms.
>
>test.owl
><http://protege-project.136.n4.nabble.com/file/n4659770/test.owl>
>
>Thanks in advance for any direction.
>
>Best regards
>
>Eddy
>
>
>
>
>--
>View this message in context:
>http://protege-project.136.n4.nabble.com/Problems-with-Hermit-reasoner-tp4
>659770.html
>Sent from the Protege-OWL 4.x Support mailing list archive at Nabble.com.
>_______________________________________________
>p4-feedback mailing list
>[hidden email]
>https://mailman.stanford.edu/mailman/listinfo/p4-feedback



------------------------------

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


End of p4-feedback Digest, Vol 88, Issue 14
*******************************************

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback