Get To The Point: Summarization with Pointer-Generator Networks
A. See, P. Liu, and C. Manning. (2017)cite arxiv:1704.04368Comment: Add METEOR evaluation results, add some citations, fix some equations (what are now equations 1, 8 and 11 were missing a bias term), fix url to pyrouge package, add acknowledgments.
Abstract
Neural sequence-to-sequence models have provided a viable new approach for
abstractive text summarization (meaning they are not restricted to simply
selecting and rearranging passages from the original text). However, these
models have two shortcomings: they are liable to reproduce factual details
inaccurately, and they tend to repeat themselves. In this work we propose a
novel architecture that augments the standard sequence-to-sequence attentional
model in two orthogonal ways. First, we use a hybrid pointer-generator network
that can copy words from the source text via pointing, which aids accurate
reproduction of information, while retaining the ability to produce novel words
through the generator. Second, we use coverage to keep track of what has been
summarized, which discourages repetition. We apply our model to the CNN / Daily
Mail summarization task, outperforming the current abstractive state-of-the-art
by at least 2 ROUGE points.
Description
[1704.04368] Get To The Point: Summarization with Pointer-Generator Networks
cite arxiv:1704.04368Comment: Add METEOR evaluation results, add some citations, fix some equations (what are now equations 1, 8 and 11 were missing a bias term), fix url to pyrouge package, add acknowledgments
%0 Generic
%1 see2017point
%A See, Abigail
%A Liu, Peter J.
%A Manning, Christopher D.
%D 2017
%K deepgeneration naacl2018 neuralnet rnn
%T Get To The Point: Summarization with Pointer-Generator Networks
%U http://arxiv.org/abs/1704.04368
%X Neural sequence-to-sequence models have provided a viable new approach for
abstractive text summarization (meaning they are not restricted to simply
selecting and rearranging passages from the original text). However, these
models have two shortcomings: they are liable to reproduce factual details
inaccurately, and they tend to repeat themselves. In this work we propose a
novel architecture that augments the standard sequence-to-sequence attentional
model in two orthogonal ways. First, we use a hybrid pointer-generator network
that can copy words from the source text via pointing, which aids accurate
reproduction of information, while retaining the ability to produce novel words
through the generator. Second, we use coverage to keep track of what has been
summarized, which discourages repetition. We apply our model to the CNN / Daily
Mail summarization task, outperforming the current abstractive state-of-the-art
by at least 2 ROUGE points.
@misc{see2017point,
abstract = {Neural sequence-to-sequence models have provided a viable new approach for
abstractive text summarization (meaning they are not restricted to simply
selecting and rearranging passages from the original text). However, these
models have two shortcomings: they are liable to reproduce factual details
inaccurately, and they tend to repeat themselves. In this work we propose a
novel architecture that augments the standard sequence-to-sequence attentional
model in two orthogonal ways. First, we use a hybrid pointer-generator network
that can copy words from the source text via pointing, which aids accurate
reproduction of information, while retaining the ability to produce novel words
through the generator. Second, we use coverage to keep track of what has been
summarized, which discourages repetition. We apply our model to the CNN / Daily
Mail summarization task, outperforming the current abstractive state-of-the-art
by at least 2 ROUGE points.},
added-at = {2018-06-01T17:41:14.000+0200},
author = {See, Abigail and Liu, Peter J. and Manning, Christopher D.},
biburl = {https://www.bibsonomy.org/bibtex/24add81f447af21750f77d3447066e3a9/albinzehe},
description = {[1704.04368] Get To The Point: Summarization with Pointer-Generator Networks},
interhash = {1d9b29f1e432f0d16e4e6b4de336f63d},
intrahash = {4add81f447af21750f77d3447066e3a9},
keywords = {deepgeneration naacl2018 neuralnet rnn},
note = {cite arxiv:1704.04368Comment: Add METEOR evaluation results, add some citations, fix some equations (what are now equations 1, 8 and 11 were missing a bias term), fix url to pyrouge package, add acknowledgments},
timestamp = {2018-06-01T17:41:14.000+0200},
title = {Get To The Point: Summarization with Pointer-Generator Networks},
url = {http://arxiv.org/abs/1704.04368},
year = 2017
}