An Overview of Natural Language Processing Models for Abstractive Text Summarization.

Abstract

Enormous amount of data is created on the internet every passing second. The International Data Corporation (IDC) projects that the total amount of digital data circulating annually around the world would sprout from 4.4 zettabytes in 2013 to 180 zettabytes in 2025. To manage this data, process it, and use it to gain valuable insight to apply in different domains; it is important to summarize long pieces of text and extract their gists in a short amount of time efficiently without much human intervention. Natural Language Processing(NLP) enabled text summarization through different models and algorithms over the years. NLP is amidst a golden era with extremely powerful models already in the public domain. This paper aims at highlighting the evolution of abstractive text summarization models from RNN based Seq2Seq models (2013) to GPT-n series models (2020).