Dilated convolutional neural network for WSD.

In language processing, the identification of correct sense of word is a difficult one until to date. The reason behind this problem is the contribution of different meanings for a single word. A single word can give multiple meanings based on their expression and the neighboring words. This necessitates research in the field of understanding the correct sense of context. The research process has been started from the golden era to modern days using different approaches like supervised learning, unsupervised and semi-supervised learning. This paper is about the supervised learning approach for the discrimination of sense of words. The approach is based on the training of a type of Convolutional neural network called Dilated Convolutional Neural Network. The practical simulation of the proposed method is implemented using WORDNET. The proposed method produced the good results with an accuracy of 98.1% on SEMEVAL TASK 11. [ABSTRACT FROM AUTHOR]