Sequence Labeling Parsing by Learning across Representations
![Thumbnail](/dspace/bitstream/handle/2183/36671/Strzyz_Michalina_2019_Sequence_labeling_parsing_learning_across_representations.pdf.jpg?sequence=5&isAllowed=y)
View/ Open
Use this link to cite
http://hdl.handle.net/2183/36671Collections
- OpenAIRE [297]
- GI-LYS - Congresos, conferencias, etc. [60]
Metadata
Show full item recordTitle
Sequence Labeling Parsing by Learning across RepresentationsDate
2019-07Citation
Michalina Strzyz, David Vilares, and Carlos Gómez-Rodríguez. 2019. Sequence Labeling Parsing by Learning across Representations. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5350–5357, Florence, Italy. Association for Computational Linguistics.
Abstract
[Absctract]: We use parsing as sequence labeling as a common framework to learn across constituency and dependency syntactic abstractions. To do so, we cast the problem as multitask learning (MTL). First, we show that adding a parsing paradigm as an auxiliary loss consistently improves the performance on the other paradigm. Secondly, we explore an MTL sequence labeling model that parses both representations, at almost no cost in terms of performance and speed. The results across the board show that on average MTL models with auxiliary losses for constituency parsing outperform single-task ones by 1.05 F1 points, and for dependency parsing by 0.62 UAS points.
Keywords
Sequence labeling
Multitask learning (MTL)
Constituency parsing
Dependency parsing
Multitask learning (MTL)
Constituency parsing
Dependency parsing
Description
The 57th Annual Meeting of the Association for Computational Linguistics (ACL) took place in Florence (Italy) at the 'Fortezza da Basso' from July 28th to August 2nd, 2019.
Editor version
Rights
Atribución 3.0 España