[expertfinder-dev] Re: Expertise and Trust Vocabs

[Axel Polleres]

As warned before, this answer only goes to the mailinglist!


I also cc:ed daniel olmedilla, since I think that

_:dani
   rdfs:seeAlso
     <http://ebiquity.umbc.edu/person/foaf/Daniel/Olmedilla/foaf.rdf>;
   xprtf:isExpertIn
     <http://en.wikipedia.org/wiki/Trust>,
     <http://en.wikipedia.org/wiki/Information_security>,

and hope he might join the discussion.

;-)


Tom Heath wrote:
> Hi Andreas, Hi all,
> 
> Just to check, did everyone who is on the ExpertFinder list receive my
> mail about my Expertise (and other things) Vocabulary? If it somehow
> got lost along the way then please let me know and do have a look at
> the archive [1]. I've had some useful feedback already but would love
> to have some more.
> 
>> Axel has mentioned that we should extend expressiveness of foaf:knows
> 
> 
> Hmm, I'm not so sure. Creating another vocabulary to complement FOAF
> by defining some additional and more expressive relations (that have
> foaf:Person as domain and range) sounds fine, but lets not try and
> overload FOAF with extra relationship semantics. We can (and should)
> do this Web-style, by linking vocabularies.

I meant to create separate vocabularies with a separate namespace,
or to link existing ones, yes.

>> I think we should try to build that on top of the trust ontology [1], [2]
>> from mindswap.
 >
> Hmm, now I'm positive that I disagree ;) ...
> 
>> I could imagine of building subproperties of different trust levels:
>> foaf:knows might be a subproperty of trust:trust0
>> foaf:loves could be a subporpety of trust:trust10
 >
> I think this would be really problematic. Firstly the modelling of
> trust:trust[x] seems like a nasty hack for several reasons:
> 
> 1) representation: can I really define my trust in someone as a 1-10
> score? The literature on questionnaire design shows that above a
> certain number people can't really distinguish between the options in
> any meaningful way, so asking someone how much they agree with
> something on a scale of e.g. 1-100 is totally futile. 1-7 is about the
> limit for meaningful ratings. On the other hand, what about if I want
> 5.5, does that need another property in the ontology? Of course if
> we're generating trust scores from some algorithm then scores are
> unlikely to be integers, and we'll need to be able to represent that.

1) I would rather only go for "boolean" ratings at all.
    You trust someone or not. Or maybe, if you want it more refined,
    you trust someone on something. But it leaves too much ambiguity to
    say I trust someone with a degree of X.

    In doubt, I would rather suggest to go for the simplest option.

    Ranks could raqther be *derived* based on "how many" trust or based
    on some other ranking function aggregating networks of trust... but
    this is not an issue for a vocabulary but rather for algorithms on
    top (ie. "enabling technologies" as we called it in previous
    discussions)

2) I'm unsure whether it is a good idea to superclass foaf:properties:
      "foaf:knows might be a subproperty of trust:trust0"
    people might have a different understanding, of what knows means.
    I could know someone and this doesn't say anything about how much or
    little I trust this person. If even, I would put it the other way
    around, but also that sounds fishy. I could trust the dalai lama,
    although I don't know him personally

> 2) querying: imagine trying to do a SPARQL query to get all the people
> from your FOAF file who you trust above average.... Imagine having to
> search for all those properties in every query rather than just saying
> trustValue > X.
 >
> Secondly, which I think is Asma's point, I think linking the semantics
> of relationships with certain trust scores is not a sound way to
> proceed, conceptually or technically. I may not love everyone I trust
> 10/10, or conversely I may not trust all my loved ones 10/10 (I love
> my fictional brother, but wouldn't lend him a penny!). One just isn't
> strongly predictive of the other.
> 
> Apologies if the response is overly strong; this is something that
> comes up from time to time and I feel really strongly about it! :)

Lively discussion is always nice, isn't it?

>> and so on...
>> first hand we won´t necessarily define all of the trust levels.
>>
>> the advantage of doing so is also that we can  use golbecks concept of
>> TopicalTrust, which means:
>>
>> Agent A trusts Agent B concerning a special Topic/Concept/Skill etc.
>>
>> Which could be used in an application where users can rate the tags of 
>> other users and express "Okay, in that topic you´re an expert"
>>
>> And finally trust levels can be calculated automatically ;-)
> 
> 
> I'm not sure if you've seen my draft Hoonoh ontology Andreas. It takes
> a slightly different approach, but fulfils the same kind of role. For
> example, using the ExpertiseRelationship class and associated
> properties one can describe an expertise relationship between a
> foaf:Person and a Topic (hoonoh:Topic, subclass of skos:Concept), and
> assign this relationship a value.

See above, I have the same concern with the numeric value here as above.
I personally wouldn't "rank" my levels of trust". At max I would make 
some comparative statements, e.g. I" I trust X more than Y" (which by 
the way I wouldn't make public, though ;-) )

>  I hope that it's a relatively clean
> bit of modelling, but would be pleased to get feedback.

My feedback for a small starting vocab would be to keep it simple
as simple as possible/necessary, as I said.

> What we're doing is generating metrics for these
> ExpertiseRelationships (and ExperienceRelationships, and
> ImpartialityRelationships) linking people to Topics, and two others
> that link people to people (AffinityRelationship and
> TrackRecordRelationship). By following this approach we can publish
> one "opinion" (that generated by our algorithms) about someone's
> expertise/experience/etc in a particular topic, but leave individual
> applications to decide whether that person is trustworthy in a
> particular context, and depending on the user needs (i.e. people's
> thresholds for trustworthiness will defer).
> 
>> URIs for various skills & tags can be provided by dbpedia

On the expertfinder wiki, we suggested that
   ACM classifications or maybe also the subject areas at the WWW
   Conferences Archive could be interesting. Also wikipedia terms (as 
used above, were a suggestion which came up several times already).
see
  http://www.rdfweb.org/topic/ExpertFinder_2fBestPractices

If somoene wants to add/comment there, feel free!!!

BTW: As a concrete exercise: What would be a good uri for 
Trust&Security? see the above example.

> Agreed. This is what we're planning, or at least topics that map to
> dbpedia if possible.
> 
> <snip/>
> 
>> @tom, when do we start the website "Sparql query of the day"?
> 
> :D As soon as possible! Would contacting SWEO be a good strategy?
> Uldis is a member of SWEO AFAIK and may be able to present the
> proposal to that group. I imagine many of the members there would be
> interested in the idea, and their contributions/backing would be very
> helpful.

not yet sure whether i understand the sparql of the day idea really, but 
sounds like fun! :-)


axel

> 
> [1] 
> http://lists.foaf-project.org/pipermail/expertfinder-dev/2007-June/000209.html 
> 
> [2] http://kmi.open.ac.uk/people/tom/tmp/draft-hoonoh-ontology.rdf
> [3] http://kmi.open.ac.uk/people/tom/tmp/sample-hoonoh-instances.rdf
> 
> 


-- 
Dr. Axel Polleres
email: axel at polleres.net  url: http://www.polleres.net/