RE: Model and sources of uncertainty

> -----Original Message-----
> From: public-xg-urw3-request@w3.org [mailto:public-xg-urw3-request@w3.org]
> On Behalf Of Paulo CG Costa
> Sent: Monday, June 18, 2007 7:23 AM
> To: public-xg-urw3@w3.org
> Subject: Re: Model and sources of uncertainty
> 
> 
> Hello Giorgos, Vipul, and MItch,
> 
> Please, find my comments embedded.
> 
> On Jun 15, 2007, at 5:01 AM, Giorgos Stoilos wrote:
> > ________________________________________
> >> From: public-xg-urw3-request@w3.org [mailto:public-xg-urw3-
> >> request@w3.org]
> >> On Behalf Of Paulo CG Costa
> >> Sent: Thursday, June 14, 2007 12:59 PM
> >> To: public-xg-urw3@w3.org
> >> Subject: Model and sources of uncertainty
> >>
> >> Dear Mitch,
> >>
> >> The model on types of uncertainty lists only three of them:
> >> 1 - Vagueness
> >> 2 - Randomness
> >> 3 - Ambiguity
> >>
> >> How about:
> >> - unreability: knowledge from a source that is not 100% trustfull,
> >> - dissonance: we see the same piece of information, but each have a
> >> distinct interpretation,
> >> - incompleteness: which is not vagueness, since you can have a
> >> clear view
> >> of just part of the information,
> >> - inconclusiveness: we have clear, deterministic, non ambiguous
> >> information, which is also complete, we both agree upon it, and
> >> the source
> >> is reliable, but it is not enough to come up with any conclusive
> >> assertion.
> >
> > I am a bit confused about some of these. How does ambiguity differ
> > from
> > dissonance?
> 
> Dissonance happens when you have distinct pieces of evidence
> supporting contradictory or conflicting views.
> If a given piece of evidence supports hypothesis H and another
> supports hypothesis notH, then we have contradiction, since H and
> notH are mutually exclusive. A Knowledge Base that has two axioms
> supporting contradictory hypothesis is inconsistent.
> Conflicting evidence is another form of dissonance, this time
> supporting hypothesis that are not mutually exclusive. Therefore, a
> Knowledge Base that has two pieces of evidence supporting conflicting
> hypothesis is not necessarily inconsistent.
> Ambiguity happens when you have one or more pieces of evidence that
> are subject to different interpretations, which can lead to distinct,
> sometimes incompatible conclusions.
> 
> In other words, dissonance (either contradiction of conflict) refers
> to the relationship between distinct pieces of evidence, whereas
> ambiguity is more related to how a given piece of evidence is
> interpreted. To explore this idea further, two pieces of evidence may
> not be ambiguous (each one is clearly supporting a specific
> hypothesis), but they might be dissonant when analyzed together.

Hmmm, clear enough, thanks! But I think that it would be better to use a
different term in place of "ambiguity", see below.

> 
> > In the definition of inconclusiveness you use the term "not
> > enough" (i.e. incomplete). Thus, is it related to incompleteness?
> 
> If you have complete information about an inherently probabilistic
> phenomena (thinking on Quantum theory might help here) then you still
> can't reach a uncertainty-free conclusion.
> 

Sooryy..; not a Quantum theory expert (not even a novice) :).

> > Some examples might also help.
> 
> Please, be aware that when it comes to devising good examples I am
> just pathetic.
> 
> Contradictory evidence:
> - Every computer in the product design department is a mac.
> - Mary works in a Dell computer from the product design department.
> Conflicting evidence:
> - People form the design department voted Mary, a brilliant software
> designer, for employee of the year.
> - John, the design department's leader, blamed Mary for the poor
> quality ratings the department got this year.
> Ambiguous evidence:
> - Sam and Max were seen in the woods, they are hunting dogs
> (ambiguity in the meaning - patent ambiguity)

So my suggestion is to use the term nonspecificity rather than ambiguity. So
in the case of Sam and Max it is not specific (the evidence is not specific
enough) what the real case is. Are they hunting those poor dogs or are they
"hunting-dogs"? The term non-specificity also captures the case of
imprecision: "Mary is in her thirties". So Mary is either 30, 31, 32,...,
39; it is not specific.

Also consider the first case (about Dell and Mac computers). One could
characterize this case as ambiguous since either every computer is a mac or
not (similar to the case of hunting dogs). The evidences are specific enough
but the knowledge is ambiguous. 

Summing up, in my view ambiguity is a superconcept of both dissonance and
non-specificity.

> - Only tall people can play Basketball (What is a tall person - lack
> of precision - latent ambiguity)

I don't think this falls into the same category. The evidence seems very
precise and not at all ambiguous to me. The fact that concept "tall" does
not have precise definition (it is fuzzy, vague, etc) is not related to the
trust, correctness, evidence, etc of the statement. Similarly, with the
concepts "brilliant" (...software designer), "poor" (...quality ratings)
found in other examples.

Greetings,
-gstoil

> 
> > BTW, I would also add "inconsistent" in the list; where in the same
> > knowledge base there exist contradictory axioms.
> 
> An inconsistent KB is the result of contradictory evidence, which is
> a special kind of dissonant evidence.
> 
> All in all, I have a slight different view with respect to Mitch's
> ontology.
> In short, if we see uncertainty as the inability to predict the
> outcome of an event, then such inability can be caused either by our
> imperfect knowledge on the event or by the nature of the event
> itself. Therefore, the only possible types of uncertainty (generally
> speaking) are:
> 
> Epistemic - due to imperfect knowledge
> Essential - due to phenomena that is intrinsically randomic
> 
> How to differentiate?
> Maybe a modified version of the Clairvoyant test:
> Suppose one has the ability to know everything about anything, and
> you ask her about the outcome of a given phenomena.  Would she be
> able to predict it?
> If yes, then we are talking about epistemic uncertainty. Else, we are
> talking about intrinsic (or any similar word) uncertainty.
> 
> Further, I'd list the possible causes (sources) that might lead to
> such situation:
> 
> 1 - You have evidence that is (are):
> Inconclusive (see above)
> Ambiguous (see above) --> one cause of ambiguous evidence is vagueness.
> Dissonant (see above)   --> encompasses contradictory and conflicting
> evidence
> Incomplete
> Unreliable
> 
> 2 - You have an essentially random event.
> 
> That is, ambiguity (for example) is not a type of uncertainty, but a
> source of uncertainty. An ambiguous statement or axiom will have many
> possible interpretations, and will thus prevent one to make
> inferences on the basis of that statement.
> 
> Anyway, the above is just an initial view that might help with the
> discussion.
> 
> All the best,
> Paulo
> 
>

Received on Monday, 18 June 2007 10:01:10 UTC