hi felipe
On 06/11/2012 03:34, Felipe Sánchez Martínez wrote:
> Dear Moses-list members,
>
> We are trying to understand how flexible it is the tree-to-* learning
> mechanism in Moses chart. Do source side annotations for an input
> sentence of the form
>
> <tree label="NP" span="0-1"/>
> <tree label="DET" span="0-0"/>
> <tree label="NOUN" span="1-1"/>
> the cat
>
> necessarily have to define a tree structure?
no, they can just be arbitrary labels on spans. No tree structure necessary
>
> Or could labels overlap in a way which is not compatible with a single
> tree structures? For instance:
>
> <tree label="NP" span="0-2"/>
> <tree label="DET" span="0-0"/>
> <tree label="N1" span="1-2"/>
> <tree label="N2" span="0-1"/>
> <tree label="ADJ" span="2-2"/>
> la bruja verde
yep. During decoding, it will either of the incompatible labels, but not
both labels.
>
> Thanks you very much in advance for your valuable help.
> Kind regards.
Note: the tree-to-* doesn't build trees on the source side, it just uses
the source labels as conditions when apply the translation rules. And,
as your questions elicited, the labels doesn't have to form a tree, any
labelling scheme will do.
Exact definition of the 'tree'-to-* decoding algorithm is in my big
writeup, section 4.2.8
http://www.statmt.org/~s0565741/ddd.pdf
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
