Deep Learning Techniques in Computational Linguistics

Abstract

Deep Learning can be defined as a subclass of machine learning whose ultimate goal was to produce an efficient technology which could change the style of current technology standards. When compared to some of the existing methods, deep learning (DL) structures have been able to perform dominantly because of its strong learning ability. The ability of DL structures increases with respect to the complicatedness of the problem or the complicated patterns within the datasets. Structure of Deep Learning architectures are also capable of extracting more beneficial features from different datasets. This survey gives an overview about deep learning and its evolution also focusing on different DL architectures, deep learning frameworks and some of the works in Natural Language Processing where deep learning is used. The paper also focuses on some of the Natural Language Processing(NLP) tasks, where deep learning methods are achieving some headway.