Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ADJs that PWN does not know #32

Closed
vcvpaiva opened this issue Jan 18, 2017 · 11 comments
Closed

ADJs that PWN does not know #32

vcvpaiva opened this issue Jan 18, 2017 · 11 comments

Comments

@vcvpaiva
Copy link
Member

vcvpaiva commented Jan 18, 2017

adjectives that are not so in PWN:
shirtless
fist
motocross
t-ball
pyramid-shaped
hooded
caribou
karate
breaded
lamp
garlic
rhino
tattoo
suit
sled
skateboard
hardwood

EDIT: removing the ones I missed before as they were considered verbs and lemmatized. don't know what to do with this kind of error "costumed, bored, excited, concentrated, laden, broken, paved, etc.."

EDIT2: after the removal of original sentences "garlic, hardwood, pyramid-shaped" disappeared

@arademaker
Copy link
Member

'roof top' pode ser typo de http://wnpt.brlcloud.com/wn/synset?id=08663774-n, não?

@arademaker
Copy link
Member

arademaker commented Jan 18, 2017

Em 'paint and costume pieces' a palavra 'costume' foi marcada como adjetivo. Quase todo substantivo pode ser adjetivo em EN certo? 'costume pieces' seria 'peças de fantasia'?

Se fosse para um novo synset, algo como:

costume adj
 wn30:derivationallyRelated 03113657-n/costume
 wn30:lexicographerFile (semantic field) adj.all
 wn30:similarTo 02468635-a/typical
 wn30:gloss 'Usually a clothing exhibiting the qualities or characteristics that identify a group or kind or category'

Mas também podemos pensar como sugestão de palavra para o synset 02468635-a que já tem typical.

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 18, 2017

@arademaker acho que a gente nao pode chamar de "typo" essas expressoes que algumas pessoas escrevem juntas, outras escrevem separado como rooftop, fistfight, "back yard", "ping pong", "down hill" (5 synsets for downhill) and "up hill". Temos varias nesse corpus, o que nos da' "hill" como adverbio, ridiculo.

e nao, acho que "Quase todo substantivo pode ser adjetivo em PT", mas nao em EN.

"costume pieces" 'e so' fantasias mesmo, uma expressao fechada. um noun-noun, eu diria.

"custom" pode ser adjetivo quando significa "personalizado/especial" "bespoke" tipo a "Trio of Costume Watches https://www.ebth.com/items/4639254-trio-of-costume-watches com TYPO

EDIT: olhei a frase "a woman wearing paint and costume pieces rides a bike on a busy street" 'e meio estranha.
mas acho que mais importante 'e o outro uso de costume como ADJ,
A costumed performer is singing and people are dancing pois temos varios desses casos onde o participio passado esta sendo usado como adjetivo mesmo, mas PWN nao vai reconhecer, pois praticamente todos os participios podem ser usados como adjetivo tanto em PT quanto em EN
um ator fantasiado esta' cantando e as pessoas dancando. #20.

@arademaker
Copy link
Member

arademaker commented Jan 18, 2017

@vcvpaiva então "costume pieces" seria uma sugestão de MWE para http://wnpt.brlcloud.com/wn/synset?id=03113835-n ? Vc está sugerindo que o parsing errou então e deveria ter marcado POS noun e uma relação de MWE em costume pieces na frase a woman wearing paint and costume pieces...

O uso de costumed como adj tem synset em http://wnpt.brlcloud.com/wn/synset?id=00455485-a . A frase foi corretamente analisada pelo SyntaxNet Parsey mas o Freeling errou POS e por isso apontou para o verbo http://wnpt.brlcloud.com/wn/synset?id=00051761-v ao inves do adj na fase de WSD.

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 18, 2017

olha acho que se "costume piece" fosse adicionado tinha que ser a

03113657-n costume | vestido, traje, roupa
(the attire worn in a play or at a fancy dress ball; "he won the prize for best costume")

que esse sim 'e fantasia. o que voce colocou e' traje tipico (de concurso de miss)

@vcvpaiva
Copy link
Member Author

e obrigada por http://wnpt.brlcloud.com/wn/synset?id=00455485-a, eu tinha procurado usando word_en, mas como usei o lema, costume, so' tinha achado verbo e substantivo. eu acho que fiz esse erro algumas vezes, vou reverificar minha lista de faltantes.

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 25, 2017

  1. shirtless should be added to PWN,
    http://www.etymonline.com/index.php?allowed_in_frame=0&search=shirtless.

  2. motocross and t-ball are sports in Wikipedia and should be nouns in PWN, I think. which means that
    A person in a motocross uniform is not wearing a helmet and riding a red motorcycle .
    motocross uniform' should be a nn, as should be motocross bike' in
    A person is looking at a motocross bike that is lying on its side and another is racing by

't-ball' should be noun and direct object in
A little girl in a pink shirt is playing t-ball and taking a swing .

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 25, 2017

caribou, garlic, lamp, rhino, hardwood, and karate are only nouns in PWN. Don't know why
The herd of caribous is crossing a road quickly
does not produce noun. (maybe the "s"?)

Similarly don't understand why
_Three boys in karate uniform are sparring _ and Three boys in karate costumes are fighting have ADJ.

lamp light in
A boy is sitting in a room and playing a piano by lamp light
and garlic bread
A man is smearing butter on a slice of garlic bread
and hardwood floor
There is no cat walking on a hardwood floor
and skateboard trick in
A woman is doing a skateboard trick on an outdoor park porch
should produce nns.

but
There is no rhino grazing in a field
is the opposite, as grazing is considered a noun, which makes rhino an adjective.

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 25, 2017

the examples with "costumed, bored, excited, concentrated, laden, broken, paved, suited, etc.." are different. The past participles are sometimes adjectives and it seems Freeling treats them as verbs, not adjectives. they do exist in PWN as adjectives, at least all the ones here.

except for
A tattooed man is on a sofa and is holding a pencil
all the adjectives from verbs are in PWN.

@vcvpaiva
Copy link
Member Author

vcvpaiva commented Jan 25, 2017

A little girl in a green coat and a boy holding a red sled are lying in the snow

is conjunction causing trouble?

@vcvpaiva
Copy link
Member Author

closing this issue as it is mixing up adjetives not in PWN, mis-lemmatized and difference of dictionaries. keep a list in issue #30

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants