Open domain relation extraction systems identify relation and argumentphrases in a sentence without relying on any underlying schema. However,current state-of-the-art relation extraction systems are available only forEnglish because of their heavy reliance on linguistic tools such aspart-of-speech taggers and dependency parsers. We present a cross-lingualannotation projection method for language independent relation extraction. Weevaluate our method on a manually annotated test set and present results onthree typologically different languages. We release these manual annotationsand extracted relations in 61 languages from Wikipedia.
展开▼
