Not sure exactly why and for what you are searching but based upon your question, if you haven’t already, I would check into Princeton’s WordNet. There are a number of people worldwide generating WordNet associations in different languages. I also believe the data Interpretation. community focuses too much on trying to use algorithms against existing difficult to find data. In my opinion more focus on the linguistic semantics, phonetics, structures etc of languages will drive to better long term understanding and interpretation solutions. Hope this helps.
Thank you, I need it to make experiment related to Arabic summarization. There is already DUC2004 dataset but the summary itself is written in English. Also there is Gigaword dataset but it is not free. Thus I'm looking for freely available dataset with Arabic summarization.