Deep Graph Convolutional Encoders for Structured Data to Text Generation

Diego  Marcheggiani; Laura Perez-Beltrachini

Deep Graph Convolutional Encoders for Structured Data to Text Generation

Diego Marcheggiani, Laura Perez-Beltrachini

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

Most previous work on neural text generation from graph-structured data relies on standard sequence-to-sequence methods. These approaches linearise the input graph to be fed to a recurrent neural network. In this paper, we propose an alternative encoder based on graph convolutional networks that directly exploits the input structure. We report results on two graphto-sequence datasets that empirically show the benefits of explicitly encoding the input graph structure.

Original language	English
Title of host publication	Proceedings of the 11th International Conference on Natural Language Generation
Place of Publication	Tilburg University, The Netherlands
Publisher	Association for Computational Linguistics
Pages	1-9
Number of pages	9
Publication status	Published - Nov 2018
Event	11th International Conference on Natural Language Generation - Tilburg, Netherlands Duration: 5 Nov 2018 → 8 Nov 2018 https://inlg2018.uvt.nl/

Conference

Conference	11th International Conference on Natural Language Generation
Abbreviated title	INLG 2018
Country/Territory	Netherlands
City	Tilburg
Period	5/11/18 → 8/11/18
Internet address	https://inlg2018.uvt.nl/

Access to Document

Deep Graph Convolutional Encoders for Structured Data to Text GenerationFinal published version, 320 KBLicence: Creative Commons: Attribution (CC-BY)

TransModal: Translating from Multiple Modalities into Text
Lapata, M.
EU government bodies
1/09/16 → 31/08/22
Project: Research

Cite this

@inproceedings{e9ca720deae44413b84ad587f372299a,

title = "Deep Graph Convolutional Encoders for Structured Data to Text Generation",

abstract = "Most previous work on neural text generation from graph-structured data relies on standard sequence-to-sequence methods. These approaches linearise the input graph to be fed to a recurrent neural network. In this paper, we propose an alternative encoder based on graph convolutional networks that directly exploits the input structure. We report results on two graphto-sequence datasets that empirically show the benefits of explicitly encoding the input graph structure.",

author = "Diego Marcheggiani and Laura Perez-Beltrachini",

year = "2018",

month = nov,

language = "English",

pages = "1--9",

booktitle = "Proceedings of the 11th International Conference on Natural Language Generation",

publisher = "Association for Computational Linguistics",

note = "11th International Conference on Natural Language Generation , INLG 2018 ; Conference date: 05-11-2018 Through 08-11-2018",

url = "https://inlg2018.uvt.nl/",

}

Marcheggiani, D & Perez-Beltrachini, L 2018, Deep Graph Convolutional Encoders for Structured Data to Text Generation. in Proceedings of the 11th International Conference on Natural Language Generation . Association for Computational Linguistics, Tilburg University, The Netherlands , pp. 1-9, 11th International Conference on Natural Language Generation , Tilburg, Netherlands, 5/11/18. <http://aclweb.org/anthology/W18-6501 >

Deep Graph Convolutional Encoders for Structured Data to Text Generation. / Marcheggiani, Diego ; Perez-Beltrachini, Laura.
Proceedings of the 11th International Conference on Natural Language Generation . Tilburg University, The Netherlands : Association for Computational Linguistics, 2018. p. 1-9.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Deep Graph Convolutional Encoders for Structured Data to Text Generation

AU - Marcheggiani, Diego

AU - Perez-Beltrachini, Laura

PY - 2018/11

Y1 - 2018/11

N2 - Most previous work on neural text generation from graph-structured data relies on standard sequence-to-sequence methods. These approaches linearise the input graph to be fed to a recurrent neural network. In this paper, we propose an alternative encoder based on graph convolutional networks that directly exploits the input structure. We report results on two graphto-sequence datasets that empirically show the benefits of explicitly encoding the input graph structure.

AB - Most previous work on neural text generation from graph-structured data relies on standard sequence-to-sequence methods. These approaches linearise the input graph to be fed to a recurrent neural network. In this paper, we propose an alternative encoder based on graph convolutional networks that directly exploits the input structure. We report results on two graphto-sequence datasets that empirically show the benefits of explicitly encoding the input graph structure.

M3 - Conference contribution

SP - 1

EP - 9

BT - Proceedings of the 11th International Conference on Natural Language Generation

PB - Association for Computational Linguistics

CY - Tilburg University, The Netherlands

T2 - 11th International Conference on Natural Language Generation

Y2 - 5 November 2018 through 8 November 2018

ER -

Deep Graph Convolutional Encoders for Structured Data to Text Generation

Abstract / Description of output

Conference

Access to Document

Fingerprint

Projects

TransModal: Translating from Multiple Modalities into Text

Cite this