Dealing with Temporalitywith SWRL (Need precisions)

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

Dealing with Temporalitywith SWRL (Need precisions)

tomekldc
Dear all,

And I read the article :
A Method for Representing and Querying Temporal Information in OWL" M. J.

O'Connor and A. K. Das. Biomedical Engineering Systems and Technologies

(Selected Papers), Springer Verlag, CCIS 127, 97-110, 2011.

I'm working on a project where I need to manage an ontology "dynamically" (I mean with time dimension).
For instance, to take Lynsie's example : Robot_A (individual) - isWorkingWith (object property) - Robot_B (individual) - TimeStamp (xsd: data type)
Knowing that I'm working with SWRL Engine, I'm still working on the last part of my "axiom" : giving a TimeStamp / DateTime to the collaboration of the two robots.

I'm still need some precisions about how (conceptually) I need to organize my ontology (the easiest way without changing all my ontology structure).

In the meanwhile I'm trying to following Michael's recommendation, but I'm not sure what he means by "Value object". Is this a class ? :
create a Value object with both a boolean
and a timestamp. In this case you no longer have the issue of needing
non-monotonic reasoning because when you update the state of your robot you
don't change the value you add a new value that says at time X the robot is
now busy/available.
Does someone have an ontology example dealing with dynamic time ?
Thank you for your time,


--

Thomak Leduc

Élève Ingénieur,
à l'École Nationale Supérieure de Cognitique



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Dealing with Temporalitywith SWRL (Need precisions)

Martin O'Connor-2

I'm still need some precisions about how (conceptually) I need to organize my ontology (the easiest way without changing all my ontology structure).

In the meanwhile I'm trying to following Michael's recommendation, but I'm not sure what he means by "Value object". Is this a class ? :
create a Value object with both a boolean
and a timestamp. In this case you no longer have the issue of needing
non-monotonic reasoning because when you update the state of your robot you
don't change the value you add a new value that says at time X the robot is
now busy/available.

If you adopt a valid-time temporal model you will need to restructure the parts of your ontology that represent entities that change over time. You will need to 'wrap' the  time-varying value into an object that contains both the value and the time instant or interval during which the value holds to be true. 

See the following for a brief description:


The valid-time temporal model is one of many possible temporal representation models but it has the advantage of being intuitive. This model was pervasive in research into extending relational database to model time, for example. There is a significant literature there.

There is a deeper question here, though (and one that is alluded to in other ongoing discussions on this mailing list): Are OWL and SWRL good technologies for modeling dynamically changing information? 

Depending on what you are doing the answer is often 'no'.

Statements in OWL and SWRL are logical statements of truth. Neither can directly model time-varying information. You will need to develop a custom approach that encodes a temporal model that will allow you to represent this type of information (such as the approach outlined in the paper you referenced). Such an approach may work nicely or may cause serious pain depending on what you are trying to do. Monotonicity and the open world assumption can often conspire to make modeling difficult.

A common solution is to use OWL and SWRL to model the time-invariant aspects of a problem and to have a separate software layer (in, for example, Java or the language of your choice) that models the dynamic aspects.
 
Martin

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Dealing with Temporalitywith SWRL (Need precisions)

Igor Toujilov-2

You can have a look at my Wiki about dynamic ontologies [1], which also addresses the monotonicity. Another article addresses the open world assumption [2].

The observer-update pattern, described at the end of [1], can be used right now to deal with time effectively (if you do not want to pollute your ontology with multitude of the time state objects):

1. Create your observer ontology;
2. Create your update ontology importing your observer ontology;
3. Load the update ontology into Protégé;
4. Change the update ontology file from outside of Protégé;
5. Protégé will ask you whether you want to reload the ontology; agree to reload;
6. Repeat from step 4.

 

I suggest a new feature in Protégé to deal with time updates: an option to suppress the pop-up box, asking to reload the ontology. Reloading the update automatically would provide a reasonable debugging environment for dynamic ontologies.

 

Cheers,

Igor

 

[1] https://sourceforge.net/p/meloproject/wiki/ontological_representation_of_procedural_programming_languages/

[2] https://sourceforge.net/p/meloproject/wiki/open_world_and_closed_world/

 

 
 
Sent: Friday, July 12, 2019 at 7:16 PM
From: "Martin O'Connor" <[hidden email]>
To: "User support for WebProtege and Protege Desktop" <[hidden email]>
Subject: Re: [protege-user] Dealing with Temporalitywith SWRL (Need precisions)
 
I'm still need some precisions about how (conceptually) I need to organize my ontology (the easiest way without changing all my ontology structure).

In the meanwhile I'm trying to following Michael's recommendation, but I'm not sure what he means by "Value object". Is this a class ? :
create a Value object with both a boolean
and a timestamp. In this case you no longer have the issue of needing
non-monotonic reasoning because when you update the state of your robot you
don't change the value you add a new value that says at time X the robot is
now busy/available.
 
If you adopt a valid-time temporal model you will need to restructure the parts of your ontology that represent entities that change over time. You will need to 'wrap' the  time-varying value into an object that contains both the value and the time instant or interval during which the value holds to be true. 
 
See the following for a brief description:
 
 
The valid-time temporal model is one of many possible temporal representation models but it has the advantage of being intuitive. This model was pervasive in research into extending relational database to model time, for example. There is a significant literature there.
 
There is a deeper question here, though (and one that is alluded to in other ongoing discussions on this mailing list): Are OWL and SWRL good technologies for modeling dynamically changing information? 
 
Depending on what you are doing the answer is often 'no'.
 
Statements in OWL and SWRL are logical statements of truth. Neither can directly model time-varying information. You will need to develop a custom approach that encodes a temporal model that will allow you to represent this type of information (such as the approach outlined in the paper you referenced). Such an approach may work nicely or may cause serious pain depending on what you are trying to do. Monotonicity and the open world assumption can often conspire to make modeling difficult.
 
A common solution is to use OWL and SWRL to model the time-invariant aspects of a problem and to have a separate software layer (in, for example, Java or the language of your choice) that models the dynamic aspects.
 
Martin
_______________________________________________ protege-user mailing list [hidden email] https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Dealing with Temporalitywith SWRL (Need precisions)

tomekldc
Hi,

Thank you all for you answers.
It's now clearer for me.

Have a nice day.



--
Sent from: http://protege-project.136.n4.nabble.com/Protege-User-f4659818.html
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user