Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies

Shuhei Kurita; Anders Søgaard

doi:10.18653/v1/P19-1232

Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies

Shuhei Kurita, Anders Søgaard

Machine Learning

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

68 Downloads (Pure)

Abstract

In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.

Original language	English
Title of host publication	Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics
Publisher	Association for Computational Linguistics
Publication date	2019
Pages	2420-2430
DOIs	https://doi.org/10.18653/v1/P19-1232
Publication status	Published - 2019
Event	57th Annual Meeting of the Association for Computational Linguistics - Florence, Italy Duration: 1 Jul 2019 → 1 Jul 2019

Conference

Conference	57th Annual Meeting of the Association for Computational Linguistics
Country/Territory	Italy
City	Florence,
Period	01/07/2019 → 01/07/2019

Access to Document

10.18653/v1/P19-1232

OA-Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First StrategiesFinal published version, 3.4 MBLicence: CC BY

Cite this

Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies. / Kurita, Shuhei; Søgaard, Anders.
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Association for Computational Linguistics, 2019. p. 2420-2430.

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

Kurita, S & Søgaard, A 2019, Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies. in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Association for Computational Linguistics, pp. 2420-2430, 57th Annual Meeting of the Association for Computational Linguistics, Florence, Italy, 01/07/2019. https://doi.org/10.18653/v1/P19-1232

@inproceedings{4fdb0f500f3f4ad1a82a97b9d154ad20,

title = "Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies",

abstract = "In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.",

author = "Shuhei Kurita and Anders S{\o}gaard",

year = "2019",

doi = "10.18653/v1/P19-1232",

language = "English",

pages = "2420--2430",

booktitle = "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics",

publisher = "Association for Computational Linguistics",

note = "57th Annual Meeting of the Association for Computational Linguistics ; Conference date: 01-07-2019 Through 01-07-2019",

}

TY - GEN

T1 - Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies

AU - Kurita, Shuhei

AU - Søgaard, Anders

PY - 2019

Y1 - 2019

N2 - In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.

AB - In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.

U2 - 10.18653/v1/P19-1232

DO - 10.18653/v1/P19-1232

M3 - Article in proceedings

SP - 2420

EP - 2430

BT - Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

PB - Association for Computational Linguistics

T2 - 57th Annual Meeting of the Association for Computational Linguistics

Y2 - 1 July 2019 through 1 July 2019

ER -