review of discussion at Karlsruhe [was: After URSW 2008]

Claudia - Thanks for getting this started.  I believe the public  
mailing list is the one that is still open, but we'll see when I hit  
send.

I had an action to write up my notes from the discussion session at  
URSW in Karlruhe.  I will add as addresses to this email the people we  
have as registered in Karlsruhe who were not on the URW3 mailing  
list.  For the most part, I have not removed duplicates in case there  
have been email address changes not reflected since the last  
maintenance of the URW3 list.  The dialog, as I captured it, is as  
follows, but feel free to offer corrections or extensions.

Trevor: The XG skimmed over the use cases we catalogued.

Peter: We should concentrate on 2 or 3 characteristic use cases for  
which there would be data we could use to develop things further.

Claudia: The work the XG did was to convince people that uncertainty  
reasoning is needed and useful, and that was accomplished.  A next  
step would be to show how to produce annotations or a different model  
for annotations.  We should use data from real problems.  If we can  
annotate using the uncertainty ontology, we should do that or replace/ 
extend the current ontology to meet the needs we see.

Umberto: Question whether something like Fuzzy OWL or other extensions  
from the DL community will suffice.  The question of capturing  
vagueness vs. probability.  The uncertainty community knows the  
formalisms to use, and it would be useful to hold tutorials to explain  
what to do and which tools can do it.  RACER implemented probabilistic  
OWL but there were problems with the model and it didn't work well.   
What we need is a standard to represent data for a given formalism.  A  
likely methodology would be to explain how to build ontologies when  
need to include uncertainty. [Kathy was not aware of an implementation  
of probabilistic OWL in RACER. Can Umberto provide a reference?]

Kathy: how to demonstrate this is the challenge.  The amniocentesis  
example is good.

Peter: back to Ora's keynote, AI symbolic logic was not very  
successful on its own but there is now more success when combining  
with non-symbolic.  Let's be pragmatic and show something works.

Ora: Symbolic methods reinforcing non-symbolic; gives some  
underpinnings to non-symbolic.  What was found was symbolic logic  
alone wasn't enough.

Ken: The discussion is good but let's look at the list of possible  
venues to continue work (slide 15 of LaskeyLaskey presentation).

-- Continue with the URSW workshop series, using it as a forum to  
discuss advances in theory and practice
-- Approach other communities, such as those dealing with health care  
and life sciences, and form a wider collaboration
–Continue research aspects
–Provide concrete problems against which to develop solutions
-- Develop a charter for and establish a new XG to work the items  
recommended by the URW3-XG
-- Investigate funding opportunities to
–formalize a dedicated effort to pursue the issues
–develop implementable solutions and tools in a reasonable time frame
All: agreed to continue URSW workshop series

Trevor: Dealing with other communities can get you lost in their  
specific problems.

Claudia: Next move is to understand our problems and our goals.

Pavel: Collaboration can be beneficial.

Kathy: Look at a challenge problem (as a way of connecting with  
another domain)?

Uwe: What we need are real data sets.

Ken: The approach could be to develop more details in how we would  
approach uncertainty representation while we investigate domains that  
can provide problems with sufficient data.  Deepen our understanding  
without getting immersed in the details of someone else's problem  
while still looking for beneficial collaboration.

[on to funding]
Peter: Good idea.

Umberto: Open project, 5 page abstract, Future Emerging Technologies.
ACTION: Umberto to look further into this.

[back to challenge problem and engaging other community]
Ken: are we ready for challenge problems?

Umberto: Tough to find area not already covered but done with  
uncertainty.

Kathy: Judged competition for use case?

Claudia: Too soon for competition.  New XG is a good idea.  Mailing  
list only is not enough.

Trevor: Ken and Kathy did a lot of work as chairs.  Will they do it  
again?

Kathy: No time while managing large project.

Ken: Currently chairing W3C task force and can provide technical  
support for running XG but not available for regular chair  
responsibilities.

Umberto:  Need to check but might be available to chair.

Kathy: Paulo might be convinced to co-chair.

Ken: Need to get ideas together for a new charter.

ACTION (to Claudia): use URW3 mailing list to begin discussion of  
charter.

ACTION (to Ken): compile and distribute notes from this discussion



Begin forwarded message:

> From: Claudia d'Amato <claudia.damato@di.uniba.it>
> Date: November 7, 2008 3:32:58 AM EST
> To: Ken Laskey <klaskey@mitre.org>, Kathryn Blackmond Laskey <klaskey@gmu.edu 
> >, Thomas Lukasiewicz <lukasiew@gmail.com>, Trevor Martin <Trevor.Martin@bristol.ac.uk 
> >, Fernando Bobillo <fbobillo@decsai.ugr.es>, Peter Vojtas <Peter.Vojtas@mff.cuni.cz 
> >, Umberto Straccia <umberto.straccia@isti.cnr.it>, Matthias Nickles  
> <M.L.Nickles@cs.bath.ac.uk>, "Paulo Cesar G. Costa"  
> <pcosta@gmu.edu>, "'N. Fanizzi'" <fanizzi@di.uniba.it>, Mitch Kokar <mkokar@vistology.com 
> >, Francis Fung <ffung@iet.com>, Pavel Smrz <smrz@fit.vutbr.cz>, gstoil@image.ntua.gr 
> , jeff.z.pan@abdn.ac.uk, vkashyap1@partners.org
> Subject: After URSW 2008
>
> Dear all,
>
> the last edition of URSW 2008 workshop is concluded. We had very  
> interesting sessions and also we have had discussions about how to  
> continue our activities. Main proposals are:
>
> - approaching a project proposal (maybe a European project). Umberto  
> Straccia will check the open possibilities
> - continuing with a new XG for other two years
>
> Regarding the second point the ideas are:
>
> - make a sketch proposal of how to tag ontologies with uncertain  
> information, focusing on methodologies
> - set up a survey of the highlighted methods (bayesian, fuzzy...)  
> for managing uncertain knowledge
> - try to give guidelines on the usage of a method w.r.t. certain  
> conditions
>
> Ken asked me to write this email. Other proposals are obviously  
> welcomed. The idea is to make a plan of our next two years  
> activities with the goal of giving more impact on our results.
>
> Regards,
>
> Claudia
>
> P.S.: it seems that the member-only mailing list is closed so I have  
> used the email address of each of you, I hope that I have forgotten  
> none. If you see someone missing please forward this email. Thanks


-----------------------------------------------------------------------------
Ken Laskey
MITRE Corporation, M/S H305      phone: 703-983-7934
7515 Colshire Drive                         fax:       703-983-1379
McLean VA 22102-7508