Spaces:

lorenzoscottb

/

phrase-entailment

like 1

Sleeping

Hi, thanks for trying the model and the feedback.

I still have to polish up the guidelines and all, but the scope of the model is not to run lexical entailment (i.e., hypernym detection). The model is trained solely to perform a very specific subset of phrase-level entailment, based on adjective-nouns phrases.

The type of question you should ask the model are limited, and should have one of three forms:

• An adjective+Noun is a Noun (e.g. A red car is a car)
• An adjective+Noun is a noun-hypernym (e.g. A red car is a vehicle)
• An adjective+Noun is a adjective+noun-hypernym (e.g. A red car is a red vehicle)

Linguistically speaking, adjectives belong to three macro classes (intersective, subsective, and intensional). From a linguistic and logical stand, these class shape the truth value of the three forms above. For instance, since red is an intersective adjective, the three from are all true. A subjective adjective like small allows just the first two, but not the last – that is, logically speaking, a small car is not a small vehicle.

Hope this clarifies the issue. 😃

I still don't get it
the potential is huge though

a fast car is a fast object - false
a fast car is a slow object - false also

so i guess i dont quite understand how it interprets the hypernymissness(if i can use that word)

Hi,

your examples are indeed interesting but again fall outside of the current scope of this specific model.

I should have been more specific in saying that, for now:

• the adjective should be the same between the two phrases (e.g. A fast car is a fast vehicle), so your example should probably be An experienced teacher is an experienced pedagogue
• the non-adjective components have to be in a hyponym-hypernym relation, and colour is not a hypernym of teacher.

This is indeed quite limiting, but the scope of the project was to eliminate clear clues that would suggest a violation of the entailment relation. The key aspect is that, in order to solve the task, the model has to focus on the compositional function that each adjective class represents. You can take a look at the paper for more details.

re: a fast car is a fast object --> False

This clearly is an error of the model, as object is (well, should be...) an hypernym of car