Title: Abstractive Summarization with Deep Learning

Abstract: Summarization aims to generate a condensed version of an input text that captures its core meaning. 
Most state-of-the-art summarization models utilize extractive approaches which are inherently limited. 
Abstractive summarization is more similar to the way of human beings generating summaries, 
but widely regarded as a difficult problem as natural language understanding and generation are involved. 
In the recent past, neural network-based sequence-to-sequence models have achieved remarkable success 
in various NLP tasks and abstractive summarization is one of them.

In this talk, I will present a few data-driven approaches to summarize a single sentence and 
to generate news headlines within the framework of encoder-decoder models.