Hello, I am trying to use paraphrase database (http://www.cis.upenn.edu/~ccb/ppdb/) to build a simplification system on Moses. Since this is not a parallel corpus, I made the phrase-table by myself using lexical and phrasal corpus and used language model made from simple Wikipedia corpus ( http://www.cs.pomona.edu/~dkauchak/simplification/). I managed to work this, but for improvement, I want to use syntactic corpus also.
So my question is: (1) Was there a better way to use paraphrase database on Moses other than just replacing the phrase-table? (2) From what corpus should I make an language model for case like this? (If you have a parallel corpus, you usually use the target corpus as a language model) (3)How can I use the syntactic corpus? Can I just add it to the phrase table? From tokawa
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
