[Pellet-users] Pronto: Probabilistic Reasoning in Pellet

[Kendall Clark]

On Oct 1, 2007, at 12:07 PM, James A Miller wrote:

>
> The issue of confidence in our data has been a concern for me from  
> day one.  The only built-in options appear to be "definitely true"  
> or "not definitely true".  When John Smith is referenced in two  
> data sources, at what point does "sameAs" make sense?  When John  
> Smith is somehow associated with a bombing, what is the likelihood  
> that he is a terrorist?  It seems that most data is not black or  
> white, and if Pronto supports shades of gray, then I would want to  
> try it.

While not our favorite use case, for various reasons, this is indeed,  
together with info integration and the bio stuff, the other big use  
case for probabilistic reasoning in Pellet via Pronto. In fact, our  
first demo of Pronto was to a Friend of Ours who works in this area.  
It does, in short, precisely what yr asking for, etc.

What we're trying to figure out now is how to release it in a way  
that allows us to find support for its continuing development, given  
that we're a small, self-financed startup.

The (non-exhaustive) possibilities include:

1. making it available via OwlSight so that people could play with it  
in the context of a web-based OWL ontology browser -- this is  
limiting in several aspects, but it's somewhat end user friendly.

2. releasing the existing code (which is command-line and API  
oriented and not really meant for end users, modelers, or domain  
experts) as GPL'd open source.

I'd love to hear from people or institutions that might be able to  
support this work to help shape our plans hereabouts.

Cheers,
Kendall
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.owldl.com/pipermail/pellet-users/attachments/20071001/0adabd6c/attachment.htm