Word Sense Induction with Attentive Context Clustering

This paper presents ACCWSI (Attentive Context Clustering WSI), a method for Word Sense Induction, suitable languages with limited resources. Pretrained on small corpus and given an ambiguous word (a query word) set of excerpts that contain it, uses attention mechanism generating context-aware embeddings, distinguishing between the different senses assigned to word. These embeddings are then clustered provide groups main common We show performs well SemEval-2 2010 WSI task. also demonstrates practical applicability shedding light meanings words in ancient languages, such as Classical Hebrew Akkadian. In near future, we intend turn into tool linguists historians.