Learning Sense Embeddings from Dictionary Definition

We introduce a method for learning to embed word senses as defined in a given set of dictionaries. In our approach, senses as definition triples, are transformed into low-dimension vectors aimed at maximizing the probability of reconstructing the definitions in an autoencoder. The method involves automatically training sense autoencoder for encoding sense definitions, automatically aligning sense definitions, and automatically generating embeddings of arbitrary description. At run-time, queries from users are mapped to the embedding space and re-ranking is performed on the sense definition retrieved. We present a prototype sense definition embedding, SenseNet, that applies the method to two dictionaries. Blind evaluation on a set of real queries shows that the method significantly outperforms a baseline based on the Lesk algorithm. Our methodology clearly supports combining multiple dictionaries resulting in additional improvement in representing sense definitions of multiple dictionaries. Although there is no distinctive header, this is the abstract. This submission template allows authors to submit their papers for review to an ACM Conference or Journal without any output design specifications incorporated at this point in the process. The ACM manuscript template is a single column document that allows authors to type their content into the pre-existing set of paragraph formatting styles applied to the sample placeholder text here. Throughout the document you will find further instructions on how to format your text.