Analyse syntaxique de langues faiblement dotées à partir de plongements de mots multilingues: Application au same du nord et au komi-zyriène

Kyung Tae Lim; Niko Partanen; Thierry Poibeau

Analyse syntaxique de langues faiblement dotées à partir de plongements de mots multilingues: Application au same du nord et au komi-zyriène

Translated title of the contribution: Syntactic analysis of under-resourced languages from multilingual word embeddings: Application to North Saami and Komi-Zyrian

Kyung Tae Lim
, Niko Partanen
, Thierry Poibeau

Dept. of Applied Artificial Intelligence

École normale supérieure

Research output: Contribution to journal › Article › peer-review

Abstract

This article presents an attempt to apply efficient parsing methods based on recursive neural networks to languages for which very few resources are available. We propose an original approach based on multilingual word embeddings acquired from different languages so as to determine the best language combination for learning. The approach yields competitive results in contexts considered as linguistically difficult.

Translated title of the contribution	Syntactic analysis of under-resourced languages from multilingual word embeddings: Application to North Saami and Komi-Zyrian
Original language	French
Pages (from-to)	67-91
Number of pages	25
Journal	Revue Traitement Automatique des Langues
Volume	59
Issue number	3
State	Published - 2018

Keywords

Komi-Zyrian
Low-resource languages
Multilingual models
North Saami
Parsing
Word embeddings

Cite this

@article{50c960685a774e24a2c447c1c1fba2ea,

title = "Analyse syntaxique de langues faiblement dot{\'e}es {\`a} partir de plongements de mots multilingues: Application au same du nord et au komi-zyri{\`e}ne",

abstract = "This article presents an attempt to apply efficient parsing methods based on recursive neural networks to languages for which very few resources are available. We propose an original approach based on multilingual word embeddings acquired from different languages so as to determine the best language combination for learning. The approach yields competitive results in contexts considered as linguistically difficult.",

keywords = "Komi-Zyrian, Low-resource languages, Multilingual models, North Saami, Parsing, Word embeddings",

author = "Lim, \{Kyung Tae\} and Niko Partanen and Thierry Poibeau",

year = "2018",

language = "French",

volume = "59",

pages = "67--91",

journal = "Revue Traitement Automatique des Langues",

issn = "1965-0906",

number = "3",

}

TY - JOUR

T1 - Analyse syntaxique de langues faiblement dotées à partir de plongements de mots multilingues

T2 - Application au same du nord et au komi-zyriène

AU - Lim, Kyung Tae

AU - Partanen, Niko

AU - Poibeau, Thierry

PY - 2018

Y1 - 2018

N2 - This article presents an attempt to apply efficient parsing methods based on recursive neural networks to languages for which very few resources are available. We propose an original approach based on multilingual word embeddings acquired from different languages so as to determine the best language combination for learning. The approach yields competitive results in contexts considered as linguistically difficult.

AB - This article presents an attempt to apply efficient parsing methods based on recursive neural networks to languages for which very few resources are available. We propose an original approach based on multilingual word embeddings acquired from different languages so as to determine the best language combination for learning. The approach yields competitive results in contexts considered as linguistically difficult.

KW - Komi-Zyrian

KW - Low-resource languages

KW - Multilingual models

KW - North Saami

KW - Parsing

KW - Word embeddings

UR - https://www.scopus.com/pages/publications/85063412201

M3 - Article

AN - SCOPUS:85063412201

SN - 1965-0906

VL - 59

SP - 67

EP - 91

JO - Revue Traitement Automatique des Langues

JF - Revue Traitement Automatique des Langues

IS - 3

ER -

Analyse syntaxique de langues faiblement dotées à partir de plongements de mots multilingues: Application au same du nord et au komi-zyriène

Abstract

Keywords

Other files and links

Fingerprint

Cite this