Multithreaded reasoner

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

Multithreaded reasoner

Aqualung
Does anybody have knowledge of an OWL DL reasoner that could take advantage
of a network of High Performance Computing (HPC) clusters--hence
dramatically speed up reasoning for big ontologies?

C

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: Multithreaded reasoner

Alan Ruttenberg-2
On Fri, Mar 5, 2010 at 11:04 AM, Cristian Cocos <[hidden email]> wrote:
> Does anybody have knowledge of an OWL DL reasoner that could take advantage
> of a network of High Performance Computing (HPC) clusters--hence
> dramatically speed up reasoning for big ontologies?

I unaware of any currently. I know that Hermit isn't even thread safe
(having asked about it recently when I was curious whether they could
parallelize the post-classification tasks such as property
classification and realization).

You could look at:

A Reasoning Broker Framework for OWL, Juergen Bock, Tuvshintur
Tserendorj, Yongchun Xu, Jens Wissmann and Stephan Grimm.
http://www.webont.org/owled/2009/papers/owled2009_submission_13.pdf

Another possibility is that the modularization work could lead to
parallelization - I don't know if people like C&P have done any work
on that. Modularization is what lets them do some kinds of incremental
reasoning (i.e. only work on modules the changes are in).

There's also a new OWLLink protocol that lets you control multiple
reasoners over XML
The OWLlink Protocol, Thorsten Liebig, Marko Luther and Olaf Noppens
http://www.owllink.org/

OWL RL is implementable in relational databases - Oracle does it. I
don't know whether the clustering that Oracle has for databases in
general would be applicable. And OWL-RL is less expressive than DL.

There is also the The Large Knowledge Collider project
http://www.larkc.eu/ - I haven't kept up with it. You could see what
they are up to, though I expect that they are concerned with less
expressive languages.

Anyways, hope these pointers help. I'm curious, of course, if others
know of different efforts.

Is there a particular ontology you are having trouble with?

-Alan
_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: Multithreaded reasoner

kotsomit
Another effort that may be worth checking is [1]. In particular Section V
discusses an OWL API-based middleware for remote communication with
reasoners through Java Serialization. This has also been successfully tested
in a multi-threaded setting, with the exception of FaCT++ due to JNI.

Dimitrios

[1] Dimitrios A. Koutsomitropoulos, Georgia D. Solomou, Tzanetos Pomonis,
Panagiotis Aggelopoulos and Theodore S. Papatheodorou. Developing
Distributed Reasoning-Based Applications for the Semantic Web. In Proc. of
the 2010 IEEE Int. Symposium on Mining and Web (MAW10), to appear.
http://www.hpclab.ceid.upatras.gr/viografika/kotsomit/pubs/maw10.pdf



> -----Original Message-----
> From: [hidden email] [mailto:p4-feedback-
> [hidden email]] On Behalf Of Alan Ruttenberg
> Sent: Friday, March 05, 2010 6:28 PM
> To: Submit feedback for Protege 4.0 beta
> Cc: Submit feedback for Protege 4.0 beta; OBI Developers
> Subject: Re: [p4-feedback] Multithreaded reasoner
>
> On Fri, Mar 5, 2010 at 11:04 AM, Cristian Cocos <[hidden email]>
> wrote:
> > Does anybody have knowledge of an OWL DL reasoner that could take
> advantage
> > of a network of High Performance Computing (HPC) clusters--hence
> > dramatically speed up reasoning for big ontologies?
>
> I unaware of any currently. I know that Hermit isn't even thread safe
> (having asked about it recently when I was curious whether they could
> parallelize the post-classification tasks such as property
> classification and realization).
>
> You could look at:
>
> A Reasoning Broker Framework for OWL, Juergen Bock, Tuvshintur
> Tserendorj, Yongchun Xu, Jens Wissmann and Stephan Grimm.
> http://www.webont.org/owled/2009/papers/owled2009_submission_13.pdf
>
> Another possibility is that the modularization work could lead to
> parallelization - I don't know if people like C&P have done any work
> on that. Modularization is what lets them do some kinds of incremental
> reasoning (i.e. only work on modules the changes are in).
>
> There's also a new OWLLink protocol that lets you control multiple
> reasoners over XML
> The OWLlink Protocol, Thorsten Liebig, Marko Luther and Olaf Noppens
> http://www.owllink.org/
>
> OWL RL is implementable in relational databases - Oracle does it. I
> don't know whether the clustering that Oracle has for databases in
> general would be applicable. And OWL-RL is less expressive than DL.
>
> There is also the The Large Knowledge Collider project
> http://www.larkc.eu/ - I haven't kept up with it. You could see what
> they are up to, though I expect that they are concerned with less
> expressive languages.
>
> Anyways, hope these pointers help. I'm curious, of course, if others
> know of different efforts.
>
> Is there a particular ontology you are having trouble with?
>
> -Alan
> _______________________________________________
> p4-feedback mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/p4-feedback
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.

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