State of the Art, Recent Developments, and Future Directions in Applying Deep Learning to Part of Speech Tagging in NLP

Abstract

Rapid information and communicationtechnology advancements have prompted widespread interest in natural language processing (NLP) applications. This has led to the development of a plethora of NLP resources. However, several obstacles stand in the way of creating reliable NLP systems that can process natural languages effectively and efficiently. Part of speech (POS) tagging is one such technology;it assigns labels to sentences or phrases inside a paragraph based on where they appear. Researchers have made great strides in POS tagging, but there is always room for improvement, especially in decreasing false positives and correctly categorizing new words. It's also important to remember that there is sure to be some confusion if you tag phrases with manypossible interpretations depending on the surrounding material. In order to effectively identify words in a particular phrase across a paragraph, POS taggers based on “deep learning (DL) and machine learning (ML)” have recently been deployed as viable solutions.