Semantic Graph Parsing with Recurrent Neural Network DAG Grammars

Federico Fancellu; Sorcha Gilroy; Adam Lopez; Maria Lapata

doi:10.18653/v1/D19-1278

Semantic Graph Parsing with Recurrent Neural Network DAG Grammars

Federico Fancellu, Sorcha Gilroy, Adam Lopez, Maria Lapata

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

Abstract

Semantic parses are directed acyclic graphs (DAGs), so semantic parsing should be modeled as graph prediction. But predicting graphs presents difficult technical challenges, so it is simpler and more common to predict the linearized graphs found in semantic parsing datasets using well-understood sequence models. The cost of this simplicity is that the predicted strings may not be wellformed graphs. We present recurrent neural network DAG grammars, a graph-aware sequence model that ensures only well-formed graphs while sidestepping many difficulties in graph prediction. We test our model on the Parallel Meaning Bank—a multilingual semantic graphbank. Our approach yields competitive results in English and establishes the first results for German, Italian and Dutch.

Original language	English
Title of host publication	Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing
Publisher	Association for Computational Linguistics
Pages	2769–2778
Number of pages	13
ISBN (Print)	978-1-950737-90-1
DOIs	https://doi.org/10.18653/v1/D19-1278
Publication status	Published - 4 Nov 2019
Event	2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing - Hong Kong, Hong Kong Duration: 3 Nov 2019 → 7 Nov 2019 https://www.emnlp-ijcnlp2019.org/

Conference

Conference	2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing
Abbreviated title	EMNLP-IJCNLP 2019
Country/Territory	Hong Kong
City	Hong Kong
Period	3/11/19 → 7/11/19
Internet address	https://www.emnlp-ijcnlp2019.org/

Access to Document

10.18653/v1/D19-1278Licence: Creative Commons: Attribution (CC-BY)

Semantic Graph Parsing_FANCELLU_DoA300819_AFVAccepted author manuscript, 181 KBLicence: Creative Commons: Attribution (CC-BY)
Semantic Graph Parsing_FANCELLU_DOA04112019_VOR_CC-BYFinal published version, 363 KBLicence: Creative Commons: Attribution (CC-BY)

https://www.aclweb.org/anthology/D19-1278/Licence: Creative Commons: Attribution (CC-BY)

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

Cite this

Fancellu, F., Gilroy, S., Lopez, A., & Lapata, M. (2019). Semantic Graph Parsing with Recurrent Neural Network DAG Grammars. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing (pp. 2769–2778). Association for Computational Linguistics. https://doi.org/10.18653/v1/D19-1278

@inproceedings{00907f1ddaec4597b0f150a5d165b78f,

title = "Semantic Graph Parsing with Recurrent Neural Network DAG Grammars",

abstract = "Semantic parses are directed acyclic graphs (DAGs), so semantic parsing should be modeled as graph prediction. But predicting graphs presents difficult technical challenges, so it is simpler and more common to predict the linearized graphs found in semantic parsing datasets using well-understood sequence models. The cost of this simplicity is that the predicted strings may not be wellformed graphs. We present recurrent neural network DAG grammars, a graph-aware sequence model that ensures only well-formed graphs while sidestepping many difficulties in graph prediction. We test our model on the Parallel Meaning Bank—a multilingual semantic graphbank. Our approach yields competitive results in English and establishes the first results for German, Italian and Dutch.",

author = "Federico Fancellu and Sorcha Gilroy and Adam Lopez and Maria Lapata",

year = "2019",

month = nov,

day = "4",

doi = "10.18653/v1/D19-1278",

language = "English",

isbn = "978-1-950737-90-1",

pages = "2769–2778",

booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing",

publisher = "Association for Computational Linguistics",

note = "2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019 ; Conference date: 03-11-2019 Through 07-11-2019",

url = "https://www.emnlp-ijcnlp2019.org/",

}

Fancellu, F, Gilroy, S, Lopez, A & Lapata, M 2019, Semantic Graph Parsing with Recurrent Neural Network DAG Grammars. in Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing. Association for Computational Linguistics, pp. 2769–2778, 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing, Hong Kong, Hong Kong, 3/11/19. https://doi.org/10.18653/v1/D19-1278

Semantic Graph Parsing with Recurrent Neural Network DAG Grammars. / Fancellu, Federico; Gilroy, Sorcha; Lopez, Adam et al.
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing. Association for Computational Linguistics, 2019. p. 2769–2778.

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

TY - GEN

T1 - Semantic Graph Parsing with Recurrent Neural Network DAG Grammars

AU - Fancellu, Federico

AU - Gilroy, Sorcha

AU - Lopez, Adam

AU - Lapata, Maria

PY - 2019/11/4

Y1 - 2019/11/4

N2 - Semantic parses are directed acyclic graphs (DAGs), so semantic parsing should be modeled as graph prediction. But predicting graphs presents difficult technical challenges, so it is simpler and more common to predict the linearized graphs found in semantic parsing datasets using well-understood sequence models. The cost of this simplicity is that the predicted strings may not be wellformed graphs. We present recurrent neural network DAG grammars, a graph-aware sequence model that ensures only well-formed graphs while sidestepping many difficulties in graph prediction. We test our model on the Parallel Meaning Bank—a multilingual semantic graphbank. Our approach yields competitive results in English and establishes the first results for German, Italian and Dutch.

AB - Semantic parses are directed acyclic graphs (DAGs), so semantic parsing should be modeled as graph prediction. But predicting graphs presents difficult technical challenges, so it is simpler and more common to predict the linearized graphs found in semantic parsing datasets using well-understood sequence models. The cost of this simplicity is that the predicted strings may not be wellformed graphs. We present recurrent neural network DAG grammars, a graph-aware sequence model that ensures only well-formed graphs while sidestepping many difficulties in graph prediction. We test our model on the Parallel Meaning Bank—a multilingual semantic graphbank. Our approach yields competitive results in English and establishes the first results for German, Italian and Dutch.

U2 - 10.18653/v1/D19-1278

DO - 10.18653/v1/D19-1278

M3 - Conference contribution

SN - 978-1-950737-90-1

SP - 2769

EP - 2778

BT - Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing

PB - Association for Computational Linguistics

T2 - 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing

Y2 - 3 November 2019 through 7 November 2019

ER -

Semantic Graph Parsing with Recurrent Neural Network DAG Grammars

Abstract

Conference

Access to Document

Fingerprint

Projects

TransModal: Translating from Multiple Modalities into Text

Cite this