Modify or retract ontology information

classic Classic list List threaded Threaded
4 messages Options
edi
Reply | Threaded
Open this post in threaded view
|

Modify or retract ontology information

edi

Hello,

 

Using Protégé we can not modify or retract any information of the ontology because of OWL and SWRL’s monotonicity. Is there any external tool to use for these purposes in Protégé environment?

 

Thank you


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

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: Modify or retract ontology information

Timothy Redmond
On 02/20/2011 01:02 AM, Edmond Jajaga wrote:

Hello,

 

Using Protégé we can not modify or retract any information of the ontology because of OWL and SWRL’s monotonicity. Is there any external tool to use for these purposes in Protégé environment?


I am not sure I understand your question.  As stated it does not appear to be describing monotonicity correctly [1].  Monotonicity is a property of the language and inference but not the tools that manipulate the language.  It essentially states that adding new information won't contradict previous conclusions. 

As an example of non-monotonic inference, there is the classical example of the statement that by default a bird can fly.  If you are then told that a Penguin is a bird then you will immediately be able to conclude that a Penguin can fly.  But you can then be told that a Penguin is an exception to this rule and cannot fly.  This additional information negates your original conclusion.  In a non-monotonic logic it does not contradict the original default that birds can fly.

It is often argued that non-monotonic logic is a better fit for how people think and is a better match for real world applications.  But non-monotonic logics tend to be much more complicated than monotonic logics.  In particular OWL and SWRL are monotonic logics.

So it is true that Protege tools cannot retract previous inferences.  But Protege is able to remove information from an ontology.  When you change the definition of a class or rearrange the class hierarchy, you are retracting axioms and replacing them with new ones.

-Timothy



[1] http://en.wikipedia.org/wiki/Monotonicity_of_entailment

 

Thank you

_______________________________________________ protege-owl mailing list [hidden email] https://mailman.stanford.edu/mailman/listinfo/protege-owl Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03


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

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
edi
Reply | Threaded
Open this post in threaded view
|

Re: Modify or retract ontology information

edi
Hi,
 
To be more clear: Certainly that Protege GUI enables fact modification and retraction, but my enquiry is when it comes to the rule layer reasoning. Because of SWRL's monotonicity we are not allowed to write any SWRL rule that would modify or retract any current information of the ontology. I am asking you how to perform these operations through SWRL or any other external tool?
These features are allowed in the W3C Recommendation - Rule Interchange Format (RIF). Espescially the RIF Production Rules Dialect can be used to model production rules [2]. Features that are notably in PRD, but not BLD, include negation and retraction of facts (thus, PRD is not monotonic).  
 
 
Thak you
 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Timothy Redmond
Sent: Tuesday, February 22, 2011 6:58 PM
To: [hidden email]
Subject: Re: [protege-owl] Modify or retract ontology information

On 02/20/2011 01:02 AM, Edmond Jajaga wrote:

Hello,

Using Protégé we can not modify or retract any information of the ontology because of OWL and SWRL’s monotonicity. Is there any external tool to use for these purposes in Protégé environment?


I am not sure I understand your question.  As stated it does not appear to be describing monotonicity correctly [1].  Monotonicity is a property of the language and inference but not the tools that manipulate the language.  It essentially states that adding new information won't contradict previous conclusions. 

As an example of non-monotonic inference, there is the classical example of the statement that by default a bird can fly.  If you are then told that a Penguin is a bird then you will immediately be able to conclude that a Penguin can fly.  But you can then be told that a Penguin is an exception to this rule and cannot fly.  This additional information negates your original conclusion.  In a non-monotonic logic it does not contradict the original default that birds can fly.

It is often argued that non-monotonic logic is a better fit for how people think and is a better match for real world applications.  But non-monotonic logics tend to be much more complicated than monotonic logics.  In particular OWL and SWRL are monotonic logics.

So it is true that Protege tools cannot retract previous inferences.  But Protege is able to remove information from an ontology.  When you change the definition of a class or rearrange the class hierarchy, you are retracting axioms and replacing them with new ones.

-Timothy



[1] http://en.wikipedia.org/wiki/Monotonicity_of_entailment

Thank you

_______________________________________________ protege-owl mailing list [hidden email] https://mailman.stanford.edu/mailman/listinfo/protege-owl Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03


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

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: Modify or retract ontology information

Thomas Russ

On Feb 23, 2011, at 12:27 AM, Edmond jajaga wrote:

>
> To be more clear: Certainly that Protege GUI enables fact  
> modification and retraction, but my enquiry is when it comes to the  
> rule layer reasoning. Because of SWRL's monotonicity we are not  
> allowed to write any SWRL rule that would modify or retract any  
> current information of the ontology. I am asking you how to perform  
> these operations through SWRL or any other external tool?

Well, you can't write them in SWRL, since that is monotonic.

Unfortunately, short of writing your own code, I don't know of any  
other external tool that will do what you want with rules, although it  
does seem that you have found some leads below:

> These features are allowed in the W3C Recommendation - Rule  
> Interchange Format (RIF). Espescially the RIF Production Rules  
> Dialect can be used to model production rules [2]. Features that are  
> notably in PRD, but not BLD, include negation and retraction of  
> facts (thus, PRD is not monotonic).
>
> [1] http://www.w3.org/2005/rules/wiki/RIF_FAQ
> [2] http://www.cse.unsw.edu.au/~billw/cs9414/notes/kr/rules/rules.html

 From looking at the implementations page from [1], it seems there may  
be some candidates.  So perhaps you could try them.  As far as I know,  
nobody has tried to integrate them with Protege, although I suppose  
that if you found a system that did what you want, you could try  
writing your own plug-in to make it available in Protege.

The only system I know of that combined a description logic with a  
production rule system is the Loom system from ISI (which I worked on).
http://www.isi.edu/isd/LOOM but that system doesn't have an OWL  
interface.  It also differs in some other ways from the OWL semantics.

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

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03