Morfessor-enriched features and multilingual training for canonical morphological segmentation

Aku Rouhe; Stig-Arne Grönroos; Sami Virpioja; Mathias Creutz; Mikko Kurimo

doi:10.18653/v1/2022.sigmorphon-1.16

Morfessor-enriched features and multilingual training for canonical morphological segmentation

Aku Rouhe, Stig-Arne Grönroos, Sami Virpioja, Mathias Creutz, Mikko Kurimo

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

Abstract

In our submission to the SIGMORPHON 2022 Shared Task on Morpheme Segmentation, we study whether an unsupervised morphological segmentation method, Morfessor, can help in a supervised setting. Previous research has shown the effectiveness of the approach in semisupervised settings with small amounts of labeled data. The current tasks vary in data size: the amount of word-level annotated training data is much larger, but the amount of sentencelevel annotated training data remains small. Our approach is to pre-segment the input data for a neural sequence-to-sequence model with the unsupervised method. As the unsupervised method can be trained with raw text data, we use Wikipedia to increase the amount of training data. In addition, we train multilingual models for the sentence-level task. The results for the Morfessor-enriched features are mixed, showing benefit for all three sentencelevel tasks but only some of the word-level tasks. The multilingual training yields considerable improvements over the monolingual sentence-level models, but it negates the effect of the enriched features.

Original language	English
Title of host publication	Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology
Editors	Garrett Nicolai, Eleanor Chodroff
Number of pages	8
Place of Publication	Stroudsburg
Publisher	The Association for Computational Linguistics
Publication date	Jun 2022
Pages	144-151
ISBN (Electronic)	978-1-955917-82-7
DOIs	https://doi.org/10.18653/v1/2022.sigmorphon-1.16
Publication status	Published - Jun 2022
MoE publication type	A4 Article in conference proceedings
Event	Workshop on Computational Research in Phonetics, Phonology, and Morphology - Seattle, United States Duration: 14 Jul 2022 → 14 Jul 2022 Conference number: 19

Fields of Science

113 Computer and information sciences
natural language processing
morphology

Access to Document

10.18653/v1/2022.sigmorphon-1.16Licence: CC BY

2022.sigmorphon-1.16Final published version, 279 KBLicence: CC BY

Cite this

Rouhe, A., Grönroos, S.-A., Virpioja, S., Creutz, M., & Kurimo, M. (2022). Morfessor-enriched features and multilingual training for canonical morphological segmentation. In G. Nicolai, & E. Chodroff (Eds.), Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology (pp. 144-151). The Association for Computational Linguistics. https://doi.org/10.18653/v1/2022.sigmorphon-1.16

Rouhe, Aku ; Grönroos, Stig-Arne ; Virpioja, Sami et al. / Morfessor-enriched features and multilingual training for canonical morphological segmentation. Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology. editor / Garrett Nicolai ; Eleanor Chodroff . Stroudsburg : The Association for Computational Linguistics, 2022. pp. 144-151

@inproceedings{1835aef2838d44fe97b63382f47d1323,

title = "Morfessor-enriched features and multilingual training for canonical morphological segmentation",

abstract = "In our submission to the SIGMORPHON 2022 Shared Task on Morpheme Segmentation, we study whether an unsupervised morphological segmentation method, Morfessor, can help in a supervised setting. Previous research has shown the effectiveness of the approach in semisupervised settings with small amounts of labeled data. The current tasks vary in data size: the amount of word-level annotated training data is much larger, but the amount of sentencelevel annotated training data remains small. Our approach is to pre-segment the input data for a neural sequence-to-sequence model with the unsupervised method. As the unsupervised method can be trained with raw text data, we use Wikipedia to increase the amount of training data. In addition, we train multilingual models for the sentence-level task. The results for the Morfessor-enriched features are mixed, showing benefit for all three sentencelevel tasks but only some of the word-level tasks. The multilingual training yields considerable improvements over the monolingual sentence-level models, but it negates the effect of the enriched features.",

keywords = "113 Computer and information sciences, natural language processing, morphology",

author = "Aku Rouhe and Stig-Arne Gr{\"o}nroos and Sami Virpioja and Mathias Creutz and Mikko Kurimo",

year = "2022",

month = jun,

doi = "10.18653/v1/2022.sigmorphon-1.16",

language = "English",

pages = "144--151",

editor = "Nicolai, {Garrett } and {Chodroff }, {Eleanor }",

booktitle = "Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology",

publisher = "The Association for Computational Linguistics",

address = "United States",

note = "Workshop on Computational Research in Phonetics, Phonology, and Morphology, SIGMORPHON ; Conference date: 14-07-2022 Through 14-07-2022",

}

Rouhe, A, Grönroos, S-A , Virpioja, S , Creutz, M & Kurimo, M 2022, Morfessor-enriched features and multilingual training for canonical morphological segmentation. in G Nicolai & E Chodroff (eds), Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology. The Association for Computational Linguistics, Stroudsburg, pp. 144-151, Workshop on Computational Research in Phonetics, Phonology, and Morphology, Seattle, Washington, United States, 14/07/2022. https://doi.org/10.18653/v1/2022.sigmorphon-1.16

Morfessor-enriched features and multilingual training for canonical morphological segmentation. / Rouhe, Aku; Grönroos, Stig-Arne ; Virpioja, Sami et al.
Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology. ed. / Garrett Nicolai; Eleanor Chodroff . Stroudsburg: The Association for Computational Linguistics, 2022. p. 144-151.

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

TY - GEN

T1 - Morfessor-enriched features and multilingual training for canonical morphological segmentation

AU - Rouhe, Aku

AU - Grönroos, Stig-Arne

AU - Virpioja, Sami

AU - Creutz, Mathias

AU - Kurimo, Mikko

N1 - Conference code: 19

PY - 2022/6

Y1 - 2022/6

N2 - In our submission to the SIGMORPHON 2022 Shared Task on Morpheme Segmentation, we study whether an unsupervised morphological segmentation method, Morfessor, can help in a supervised setting. Previous research has shown the effectiveness of the approach in semisupervised settings with small amounts of labeled data. The current tasks vary in data size: the amount of word-level annotated training data is much larger, but the amount of sentencelevel annotated training data remains small. Our approach is to pre-segment the input data for a neural sequence-to-sequence model with the unsupervised method. As the unsupervised method can be trained with raw text data, we use Wikipedia to increase the amount of training data. In addition, we train multilingual models for the sentence-level task. The results for the Morfessor-enriched features are mixed, showing benefit for all three sentencelevel tasks but only some of the word-level tasks. The multilingual training yields considerable improvements over the monolingual sentence-level models, but it negates the effect of the enriched features.

AB - In our submission to the SIGMORPHON 2022 Shared Task on Morpheme Segmentation, we study whether an unsupervised morphological segmentation method, Morfessor, can help in a supervised setting. Previous research has shown the effectiveness of the approach in semisupervised settings with small amounts of labeled data. The current tasks vary in data size: the amount of word-level annotated training data is much larger, but the amount of sentencelevel annotated training data remains small. Our approach is to pre-segment the input data for a neural sequence-to-sequence model with the unsupervised method. As the unsupervised method can be trained with raw text data, we use Wikipedia to increase the amount of training data. In addition, we train multilingual models for the sentence-level task. The results for the Morfessor-enriched features are mixed, showing benefit for all three sentencelevel tasks but only some of the word-level tasks. The multilingual training yields considerable improvements over the monolingual sentence-level models, but it negates the effect of the enriched features.

KW - 113 Computer and information sciences

KW - natural language processing

KW - morphology

U2 - 10.18653/v1/2022.sigmorphon-1.16

DO - 10.18653/v1/2022.sigmorphon-1.16

M3 - Conference contribution

SP - 144

EP - 151

BT - Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology

A2 - Nicolai, Garrett

A2 - Chodroff , Eleanor

PB - The Association for Computational Linguistics

CY - Stroudsburg

T2 - Workshop on Computational Research in Phonetics, Phonology, and Morphology

Y2 - 14 July 2022 through 14 July 2022

ER -

Rouhe A, Grönroos SA , Virpioja S , Creutz M, Kurimo M. Morfessor-enriched features and multilingual training for canonical morphological segmentation. In Nicolai G, Chodroff E, editors, Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology. Stroudsburg: The Association for Computational Linguistics. 2022. p. 144-151 doi: 10.18653/v1/2022.sigmorphon-1.16