Graph-to-sequence learning using Gated Graph Neural Networks

Daniel Beck; Gholamreza Haffari; Trevor Cohn

Graph-to-sequence learning using Gated Graph Neural Networks

Daniel Beck, Gholamreza Haffari, Trevor Cohn

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

Abstract

Many NLP applications can be framed as a graph-to-sequence learning problem.
Previous work proposing neural architectures on this setting obtained promising
results compared to grammar-based approaches but still rely on linearisation
heuristics and/or standard recurrent networks to achieve the best performance.
In this work, we propose a new model that encodes the full structural information contained in the graph. Our architecture couples the recently proposed Gated Graph Neural Networks with an input transformation that allows nodes and edges to have their own hidden representations, while tackling the parameter explosion problem present in previous work. Experimental results show that our model outperforms strong baselines in generation from AMR graphs and syntax-based neural machine translation.

Original language	English
Title of host publication	ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics
Subtitle of host publication	Proceedings of the Conference, Vol. 1 (Long Papers)
Editors	Iryna Gurevych, Yusuke Miyao
Place of Publication	Stroudsburg PA USA
Publisher	Association for Computational Linguistics (ACL)
Pages	273-283
Number of pages	11
ISBN (Print)	9781948087322
Publication status	Published - 2018
Event	Annual Meeting of the Association of Computational Linguistics 2018 - Melbourne, Australia Duration: 15 Jul 2018 → 20 Jul 2018 Conference number: 56th https://aclanthology.info/events/acl-2018

Conference

Conference	Annual Meeting of the Association of Computational Linguistics 2018
Abbreviated title	ACL 2018
Country/Territory	Australia
City	Melbourne
Period	15/07/18 → 20/07/18
Internet address	https://aclanthology.info/events/acl-2018

Access to Document

247469562-oaFinal published version, 310 KB

http://aclweb.org/anthology/P18-1026

Cite this

Beck, D., Haffari, G., & Cohn, T. (2018). Graph-to-sequence learning using Gated Graph Neural Networks. In I. Gurevych, & Y. Miyao (Eds.), ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 1 (Long Papers) (pp. 273-283). Association for Computational Linguistics (ACL). http://aclweb.org/anthology/P18-1026

Beck, Daniel ; Haffari, Gholamreza ; Cohn, Trevor. / Graph-to-sequence learning using Gated Graph Neural Networks. ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 1 (Long Papers). editor / Iryna Gurevych ; Yusuke Miyao. Stroudsburg PA USA : Association for Computational Linguistics (ACL), 2018. pp. 273-283

@inproceedings{90477fdff705421b81b87afbd543cbd5,

title = "Graph-to-sequence learning using Gated Graph Neural Networks",

abstract = "Many NLP applications can be framed as a graph-to-sequence learning problem.Previous work proposing neural architectures on this setting obtained promisingresults compared to grammar-based approaches but still rely on linearisationheuristics and/or standard recurrent networks to achieve the best performance.In this work, we propose a new model that encodes the full structural information contained in the graph. Our architecture couples the recently proposed Gated Graph Neural Networks with an input transformation that allows nodes and edges to have their own hidden representations, while tackling the parameter explosion problem present in previous work. Experimental results show that our model outperforms strong baselines in generation from AMR graphs and syntax-based neural machine translation.",

author = "Daniel Beck and Gholamreza Haffari and Trevor Cohn",

year = "2018",

language = "English",

isbn = "9781948087322",

pages = "273--283",

editor = "Iryna Gurevych and Yusuke Miyao",

booktitle = "ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics",

publisher = "Association for Computational Linguistics (ACL)",

note = "Annual Meeting of the Association of Computational Linguistics 2018, ACL 2018 ; Conference date: 15-07-2018 Through 20-07-2018",

url = "https://aclanthology.info/events/acl-2018",

}

Beck, D, Haffari, G & Cohn, T 2018, Graph-to-sequence learning using Gated Graph Neural Networks. in I Gurevych & Y Miyao (eds), ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 1 (Long Papers). Association for Computational Linguistics (ACL), Stroudsburg PA USA, pp. 273-283, Annual Meeting of the Association of Computational Linguistics 2018, Melbourne, Victoria, Australia, 15/07/18. <http://aclweb.org/anthology/P18-1026>

Graph-to-sequence learning using Gated Graph Neural Networks. / Beck, Daniel; Haffari, Gholamreza; Cohn, Trevor.
ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 1 (Long Papers). ed. / Iryna Gurevych; Yusuke Miyao. Stroudsburg PA USA: Association for Computational Linguistics (ACL), 2018. p. 273-283.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Graph-to-sequence learning using Gated Graph Neural Networks

AU - Beck, Daniel

AU - Haffari, Gholamreza

AU - Cohn, Trevor

N1 - Conference code: 56th

PY - 2018

Y1 - 2018

N2 - Many NLP applications can be framed as a graph-to-sequence learning problem.Previous work proposing neural architectures on this setting obtained promisingresults compared to grammar-based approaches but still rely on linearisationheuristics and/or standard recurrent networks to achieve the best performance.In this work, we propose a new model that encodes the full structural information contained in the graph. Our architecture couples the recently proposed Gated Graph Neural Networks with an input transformation that allows nodes and edges to have their own hidden representations, while tackling the parameter explosion problem present in previous work. Experimental results show that our model outperforms strong baselines in generation from AMR graphs and syntax-based neural machine translation.

AB - Many NLP applications can be framed as a graph-to-sequence learning problem.Previous work proposing neural architectures on this setting obtained promisingresults compared to grammar-based approaches but still rely on linearisationheuristics and/or standard recurrent networks to achieve the best performance.In this work, we propose a new model that encodes the full structural information contained in the graph. Our architecture couples the recently proposed Gated Graph Neural Networks with an input transformation that allows nodes and edges to have their own hidden representations, while tackling the parameter explosion problem present in previous work. Experimental results show that our model outperforms strong baselines in generation from AMR graphs and syntax-based neural machine translation.

M3 - Conference Paper

SN - 9781948087322

SP - 273

EP - 283

BT - ACL 2018 - The 56th Annual Meeting of the Association for Computational Linguistics

A2 - Gurevych, Iryna

A2 - Miyao, Yusuke

PB - Association for Computational Linguistics (ACL)

CY - Stroudsburg PA USA

T2 - Annual Meeting of the Association of Computational Linguistics 2018

Y2 - 15 July 2018 through 20 July 2018

ER -