Structured Neural Summarization

Patrick Fernandes, Miltiadis Allamanis, Marc Brockschmidt
11/5/2018
cs.LGcs.CLcs.SE

Official code: CoderPat/structured-neural-summarization (72 stars, MIT, last commit 2019-05-09).

Abstract

Summarization of long sequences into a concise statement is a core problem in natural language processing, requiring non-trivial understanding of the input. Based on the promising results of graph neural networks on highly structured data, we develop a framework to extend existing sequence encoders with a graph component that can reason about long-distance relationships in weakly structured data such as text. In an extensive evaluation, we show that the resulting hybrid sequence-graph models outperform both pure sequence models as well as pure graph models on a range of summarization tasks.

View on arXivView PDF

Code Implementations(1)

CoderPat/structured-neural-summarizationOfficial100%
MIT

A repository with the code for the paper with the same title

7226Java, ShellMar 12, 20197 years agoMIT

Similar Papers

Best of Both Worlds: Advantages of Hybrid Graph Sequence Models

84% match1 repo2024

HiPool: Modeling Long Documents Using Graph Neural Networks

83% match1 repo222023

Deep Graph Convolutional Encoders for Structured Data to Text Generation

83% match1 repo1522018

Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks

83% match1 repo8922015

Talk like a Graph: Encoding Graphs for Large Language Models

82% match3 repos39.0k2023

GraphIE: A Graph-Based Framework for Information Extraction

82% match1 repo1102018

Cite this paper

@article{fernandes2018structured,
  title  = {Structured Neural Summarization},
  author = {Patrick Fernandes and Miltiadis Allamanis and Marc Brockschmidt},
  year   = {2018},
  eprint = {1811.01824},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1811.01824}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1811.01824.svg)](https://codeofpaper.com/paper/1811.01824)

Discussion