From ILI code to human language

[Daniel]

Hi All,

I am trying to replicate UKB's good results for Spanish.

I have UKB working on macos following the readme from the github repo, and I downloaded Wordnet 3.0 ILI graphs available at the IXA website.

I want to disambiguate the word "planta" in a given context.

From the dictionary, I see that "planta" has 6 different ILI codes.

1) ili-30-03956922-n 

2) ili-30-00017222-n 

3) ili-30-03365991-n 

4) ili-30-03316406-n 

5) ili-30-03367147-n 

6) ili-30-05577190-n

and when I run:

./ukb_wsd --ppr -K wnet30g_rels.bin -D wn30sp.lex context.txt

I get:

cxt1 w3  ili-30-03316406-n !! planta

Everything looks fine so far, but now how do I get from that particular ILI code to human language. Any guidance will be of help.

Thanks!

D.