W3C home > Mailing lists > Public > public-ontolex@w3.org > February 2013

Re: Next Ontolex Meeting, tomorrow, Feb 22, 12:00

From: John McCrae <jmccrae@cit-ec.uni-bielefeld.de>
Date: Thu, 21 Feb 2013 19:04:57 +0100
Message-ID: <CAC5njqrmcvvi1ZApPgCkS3ZfrtuZMgU-SF9AHkigZOH+C3L2mg@mail.gmail.com>
To: Paul Buitelaar <paul.buitelaar@deri.org>
Cc: public-ontolex@w3.org
Hi,

In lemon we have previously said that lexica are monolingual to avoid
having to repeat the language for each entry. Also I think there might be a
bit of an error here

:lexicon lemon:entry :left ;
     lemon:language "en" .

Is the same as:

:lexicon lemon:entry :left .
:lexicon lemon:language "en" .

I guess you meant to put the language on the entry?

I did find something that really puzzled me:

:left upper lobe posterior segment artery:decomposition (
    [ lemon:element :artery ]
    [ lemon:element :left upper lobe artery ]
    [ lemon:element :posterior segment artery ] ) .

According to the existing *lemon* documentation we use decomposition to
indicate tokenization or decompounding... hence you are saying the
tokenization of "left upper lobe posterior segment arterty" is
["artery","left upper lobe artery","posterior segment artery"]. I'm pretty
sure what you mean to use is the subterm property as we have already
discussed several times.

However I am still puzzled as to what you are attempting to represent in
the lexicon. These terms have clear subclass or partOf relations that can
be represented in the ontology, which is clearly of use. However the
lexical relationships are fairly trivial and could be easily constructed by
examining the tokenization of each of these terms. I guess you want to
represent that

X Y artery = X artery + Y artery

On the lexical layer, this is trivial... we need only compare the lexical
entries of the tokens to establish this. A single property makes this more
explicit (and hence easier to query), but I don't feel there would be any
need for extra modelling beyond the standard syntactic analysis of lexical
entries, as such relationships are easy to infer.

On the ontological layer, we have that XYArtery = XArtery ∏ YArtery, which
can be represented in OWL

My only guess is that you want to do some sense level mapping X Y artery
uses the words "X" exactly as "X artery" does (i.e., with the same sense)
and "Y" exactly as "Y artery" does. For this the modelling would be complex
and I don't clearly see the application, perhaps you could elucidate?

Regards,
John

On Thu, Feb 21, 2013 at 5:11 PM, Paul Buitelaar <paul.buitelaar@deri.org>wrote:

> On 21/02/2013 16:07, Philipp Cimiano wrote:
>
>> And you do not need to repeat n times that the lexicon is English:
>>
>
> that answers my question - i.e. if we mix lexical entries from different
> languages then we need to add the language ID for each entry - if not then
> not as the language ID will be a metadatum on the lexicon
>
>
>
>> :lexicon lemon:entry :left ;
>>         lemon:language "en" .
>>
>> :lexicon lemon:entry :upper ;
>>         lemon:language "en" .
>>
>> :lexicon lemon:entry :lobe ;
>>         lemon:language "en" .
>>
>> :lexicon lemon:entry :posterior ;
>>         lemon:language "en" .
>>
>> :lexicon lemon:entry :segment ;
>>         lemon:language "en" .
>>
>> :lexicon lemon:entry :artery
>>         lemon:language "en" .
>>
>> ;-)
>>
>>
>>
>>
>> Am 21.02.13 16:55, schrieb Paul Buitelaar:
>>
>>> Hi Philipp, we added lemon RDF to the examples on
>>>
>>> http://www.w3.org/community/**ontolex/wiki/Specification_of_**
>>> Requirements_on_**Terminological_Analysis<http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements_on_Terminological_Analysis>
>>>
>>>
>>> Cheers
>>>
>>>
>>> Paul
>>>
>>> On 21/02/2013 07:28, Philipp Cimiano wrote:
>>>
>>>> Dear all,
>>>>
>>>> this is to inform you that tomorrow we will have our regular ontolex
>>>>  teleconference.
>>>>
>>>> We will continue our discussion of particular examples, in particular
>>>>  looking at the contributions here:
>>>>
>>>> http://www.w3.org/community/**ontolex/wiki/Specification_of_**
>>>> Requirements/Lexicon-Ontology-**Mapping<http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements/Lexicon-Ontology-Mapping>
>>>>
>>>>
>>>>
>>>>
>>>> http://www.w3.org/community/**ontolex/wiki/Specification_of_**
>>>> Requirements_on_**Terminological_Analysis<http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements_on_Terminological_Analysis>
>>>>
>>>>
>>>>
>>>>
>>>> http://www.w3.org/community/**ontolex/wiki/Specification_of_**
>>>> Requirements/Properties-and-**Relations-of-Entries<http://www.w3.org/community/ontolex/wiki/Specification_of_Requirements/Properties-and-Relations-of-Entries>
>>>>
>>>>
>>>>
>>>>
>>>> Paul/Mihael: I do not see any RDF examples in the wiki yet, will you
>>>>  manage to work them out until tomorrow?
>>>>
>>>> Elena/Lupe: same for you, I do not see any RDF examples in the wiki
>>>> yet
>>>>
>>>> John: you wanted to work out how to represent Lexical Nets* in lemon,
>>>>  with WordNet as an example; can you please provide some RDF code for
>>>> this?
>>>>
>>>> From then on, I would discuss how to i) model lexico-syntactic
>>>> patterns (Dagmar: are you going to be there?) and ii) how to link
>>>> entries across resources. We could start with Framenet / Verbnet for
>>>> example, then showing how they can be used within a lemon lexical
>>>> entry. Linking to wiktionary could be also considered as well as to
>>>> ISOCAT.
>>>>
>>>> Another issue will be to look at modelling terminological resources
>>>> in the lexicon-ontology model.
>>>>
>>>> Thanks to everybody. We have to get concrete and produce some (draft
>>>> of a) spec this year ;-)
>>>>
>>>> Philipp.
>>>>
>>>>
>>>
>>>
>>
>> --
>> Prof. Dr. Philipp Cimiano
>> Semantic Computing Group
>> Excellence Cluster - Cognitive Interaction Technology (CITEC)
>> University of Bielefeld
>>
>> Phone: +49 521 106 12249
>> Fax: +49 521 106 12412
>> Mail:cimiano@cit-ec.uni-**bielefeld.de<Mail%3Acimiano@cit-ec.uni-bielefeld.de>
>>
>> Room H-127
>> Morgenbreede 39
>> 33615 Bielefeld
>>
>>
>
>
Received on Thursday, 21 February 2013 18:05:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 21 February 2013 18:05:25 GMT