Hi Marcin
I was actually thinking for something even more abstract. To adjust
your example:
or
.*xxx.*
In this case LT core would only know that there's set of extra
"dictionary" information that has category and a value. This way we
can add many different type of dictionaries, that could be specific to
each language and rules for that language will use that
language-specific information.
Core LT will allow each language to provide this extra information and
will not care much about its content.
Regards,
Andriy
2016-02-02 14:40 GMT-05:00 Marcin Miłkowski :
> W dniu 02.02.2016 o 18:08, Andriy Rysin pisze:
>> Hey Marcin
>>
>> this is great addition, though I have one remark. Besides valency
>> information some other type of information could be useful too (if we
>> starting to head this direction). E.g. I have rules in Ukrainian that
>> suggests superlative form for adjective when "самий" (very) + base
>> form is used. Currently I have the relation between base form and
>> comparative/superlative forms encoded in the dictionary but in general
>> this is higher-level information that should be stored outside of the
>> tag dictionary.
>
> I would argue that in some languages (at least in Polish and English)
> this is not a semantic-level information, this is a grammatical
> information, or morphosyntactic information.
>
>>
>> I am wondering if we could develop more generic approach for such
>> additional (semantic) information, e.g. split each type of this info
>> into category and allow generic references in the token/exception,
>> something like this:
>>
>> > semantic_info=":"/>
>>
>> or even as a subelement (I assume semantic information can get pretty
>> long/complicated so child element may be better choice and will allow
>> to add new attributes easily on it later)
>>
>>
>> > value=""/>
>> > value=""/>
>>
>>
>> so in valency case you described (1st case) it could be:
>>
>>
>>
>>
>
> Valency is definitely not a semantic category:
>
> https://en.wikipedia.org/wiki/Valency_(linguistics)
>
> But your approach seems quite elegant. I would argue that valency is one
> kind of information that should be treated as key-value
>
>
>
>
>
> This would match a verb that takes an accusative noun phrase (of course,
> the values would be defined per valency lexicon in a language). There
> are free valency lexicons for many languages beside Polish.
>
>>
>> Thus if we add other semantic information into LT we can use this info
>> in the logic without changing the LT core.
>
> The core XML parsing will have to be changed anyway.
>
> Best,
> Marcin
>
>>
>> Thanks
>> Andriy
>>
>> 2016-01-28 7:30 GMT-05:00 Marcin Miłkowski :
>>> Hi all,
>>>
>>> To allow for better disambiguation and have better rules, I need to
>>> include a valency dictionary with LT. These are dictionaries that
>>> specify which grammatical cases or prepositions go with which verbs etc.
>>> There are such resources for many languages that we support. And using
>>> these resources, we could enrich POS tag disambiguation a lot (I'm using
>>> a horribly long regular expression right now instead of a dictionary,
>>> for example), and write up a lot of important rules.
>>>
>>> The obvious choice for representing the dictionary (which is available
>>> for Polish on a fairly liberal license) is to use a finite-state lexicon
>>> that we normally use for taggers. The dictionary will be applied after
>>> tagging because valency dictionary will require POS tag + lexeme
>>> information. In Polish, the entries look like this:
>>>
>>> absurdalny: pewny: : : : {prepnp(dla,gen)}
>>> absurdalny: pewny: : : : {prepnp(w,loc)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(gdy)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(int)}
>>> absurdalny: potoczny: : pred: : {prepnp(dla,gen)}+{cp(jak)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(jeśli)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(kiedy)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(że)}
>>> absurdalny: pewny: : pred: : {prepnp(dla,gen)}+{cp(żeby)}
>>>
>>> But for French (see http://bach.arts.kuleuven.be/dicovalence/) they are
>>> paragraph-based:
>>>
>>> VAL$abaisser: P0 P1
>>> VTYPE$ predicator simple
>>> VERB$ ABAISSER/abaisser
>>> NUM$10
>>> EG$ il faudra abaisser la persienne
>>> TR_DU$ laten zakken, neerhalen, neerlaten, doen dalen
>>> TR_EN$ let down, lower
>>> FRAME$ subj:pron|n:[hum], obj:pron|n:[nhum,?abs]
>>> P0$ (que), qui, je, nous, elle, il, ils, on, (ça), (ceci), celui-ci,
>>> ceux-ci
>>> P1$ que, la, le, les, en Q, ça, ceci, celui-ci, ceux-ci
>>> RP$ passif être, se passif
>>> AUX$avoir
>>>
>>>
>>> VAL$abaisser: P0 (P1)
>>> VTYPE$ predicator simple
>>> VERB$ ABAISSER/abaisser
>>> NUM$20
>>> EG$ il a raconté cette anecdote pour m'abaisser
>>> TR_DU$ vernederen, kleineren
>>> TR_EN$ humiliate
>>> FRAME$ subj:pron|n:[hum], ?obj:pron|n:[hum]
>>> P0$ (que), qu
