Apparently, the compoents of the name node are compared to familiar phrases (I guess) and changed in the graph compared to the sentence.
For instance:
The sentence is: Who is the owner of Universal Studios?
The graph is:
'# ::snt Who is the owner of Universal Studios?
(o / own-01
:arg0 (a / amr-unknown)
:arg1 (c / company
:name (n / name
:op1 "Universal"
:op2 "Sstudio")))'
"Studios" is changed to "sStudio".
Another example:
The sentence is: When did Dracula's creator die?
The graph is:
(d / die-01
:ARG1 (p / person
:ARG0-of (c / create-01
:ARG1 (b / book
:name (n / name
:op1 "drago"))))
How can I prevent the model to change the strings? Especially in the name nodes...
Apparently, the compoents of the name node are compared to familiar phrases (I guess) and changed in the graph compared to the sentence.
For instance: The sentence is: Who is the owner of Universal Studios? The graph is: '# ::snt Who is the owner of Universal Studios? (o / own-01 :arg0 (a / amr-unknown) :arg1 (c / company :name (n / name :op1 "Universal" :op2 "Sstudio")))'
"Studios" is changed to "sStudio".
Another example: The sentence is: When did Dracula's creator die? The graph is: (d / die-01 :ARG1 (p / person :ARG0-of (c / create-01 :ARG1 (b / book :name (n / name :op1 "drago"))))
How can I prevent the model to change the strings? Especially in the name nodes...