Computing semantic similarity from bilingual dictionaries

In this paper, we address the task of calculating monoand bilingual semantic similarity. We introduce a method that, in order to arrive at a measure of semantic relatedness, exploits the information implicitly contained in bilingual dictionaries. Through experiments we show that our method performs well, with a performance comparable to approaches based on hierarchical knowledge bases and corpus statistics. The advantage of our approach is that it solely relies on easily available bilingual dictionaries and that it is capable of computing monoand bilingual semantic relatedness at the same time.