# Re: [URW3 public] Re: [URW3] ... three questions based on the last telecon

Date: Wed, 18 Jul 2007 07:43:48 -0400
To: <mkokar@vistology.com>
```
Again, I think this implies that a probability value alone is
insufficient because you need to know something about how the value was
assigned before you can combine it with other probabilities, i.e.
values that may derive from inconsistent approaches.

Ken

On Jul 18, 2007, at 6:50 AM, Mitch Kokar wrote:

> This is a good point. So in one of my previous emails I suggested that
> we
> have "complex sentences", i.e., sentences consisting of multiple
> sentences.
> Then each sentence could have uncertainty assigned to it. I still
> think an
> "elementary sentence" could be assigned a probability. For complex
> sentences
> we could have probabilities for the sentences that are "part of" the
> sentence, as well as the overall probability for the complex sentence.
> This
> would provide the flexibility that you are asking for, right?
>
> ==Mitch
>
>
>
>> -----Original Message-----
>> From: public-xg-urw3-request@w3.org
>> [mailto:public-xg-urw3-request@w3.org] On Behalf Of
>> Sent: Tuesday, July 17, 2007 11:31 PM
>> Cc: public-xg-urw3@w3.org
>> Subject: Re: [URW3 public] Re: [URW3] ... three questions
>> based on the last telecon
>>
>>
>>>> What the sentence is about is important for our decision about
>>>> uncertainty assignment - e.g. if I know a contradicting
>> information,
>>>> or a consequence from a trusted site, it will influence
>> my uncertainty
>>>> assignment.
>>>> Uncertainty about the weather is no more uncertain when
>> the tome is
>>>> gone
>>>
>>> So it appears that I may need to convey information on
>> what influenced
>>> my uncertainty assignment.  Note, this is not saying I need to
>>> represent what the sentence is about but rather I may
>> need to point to
>>> the mechanisms that were developed by "other communities"
>> and that I am
>>> using as the (or a) basis of my assessment.
>>
>> I note here that probability is not truth-functional.
>> That is, if I know the probability of A and the
>> probability of B, I don't necessarily know the probability
>> of A-and-B.  This is a very important characteristic of
>> probability. It is a source of great power, and it is also
>> the reason straightforward attempts to do uncertainty
>> propagation by attaching "certainty factors" to
>> propositions and rules works only in very constrained
>> problems.  This is important for us, because annotating
>> sentences with uncertainty values isn't going to work for
>> many interesting problems.
>>
>> Kathy
>>
>>
>
>
>
------------------------------------------------------------------------
-----