Symposium on Matching and Meaning Workshop on Matching and Meaning Programme Chairs

In this paper, we present a method for assigning examplesentences to each sense of words in WordNet. The key idea is thatthe method assigns each sense of a word w collected from not onlythe sentences containing w, but also sentences containing words withsemantically related to w. Because a collection of a context of thetarget word w is a similar syntactic behavior, even collected from avery very large corpora. The evaluation result showed that example-assignment based on groups of similar words significantly improvedthe retrieval of context for word sense, which helps to determine theexact definition of polysemous words.