Abstractive Text Summarization Using Sequence-to-Sequence RNNs and Beyond

TL;DR

This paper advances abstractive text summarization by developing novel sequence-to-sequence RNN models with attention, addressing key challenges like keyword modeling and rare word generation, and introduces a new multi-sentence summary dataset with benchmarks.

Contribution

The paper introduces several innovative models that improve summarization quality by handling key-words, hierarchy, and unseen words, along with a new dataset and performance benchmarks.

Findings

Achieved state-of-the-art results on two corpora.

Proposed models improve handling of rare and unseen words.

Established new benchmarks for multi-sentence summarization.

Abstract

In this work, we model abstractive text summarization using Attentional Encoder-Decoder Recurrent Neural Networks, and show that they achieve state-of-the-art performance on two different corpora. We propose several novel models that address critical problems in summarization that are not adequately modeled by the basic architecture, such as modeling key-words, capturing the hierarchy of sentence-to-word structure, and emitting words that are rare or unseen at training time. Our work shows that many of our proposed models contribute to further improvement in performance. We also propose a new dataset consisting of multi-sentence summaries, and establish performance benchmarks for further research.