Neural disambiguation of lemma and part of speech in morphologically rich languages

José María Hoya Quecedo; Maximilian W. Koppatz; Giacomo Furlan; Roman Yangarber

Neural disambiguation of lemma and part of speech in morphologically rich languages

José María Hoya Quecedo, Maximilian W. Koppatz, Giacomo Furlan, Roman Yangarber

Forskningsoutput: Övriga bidrag

Sammanfattning

We consider the problem of disambiguating the lemma and part of speech of ambiguous words in morphologically rich languages. We propose a method for disambiguating ambiguous words in context, using a large un-annotated corpus of text, and a morphological analyser -- with no manual disambiguation or data annotation. We assume that the morphological analyser produces multiple analyses for ambiguous words. The idea is to train recurrent neural networks on the output that the morphological analyser produces for unambiguous words. We present performance on POS and lemma disambiguation that reaches or surpasses the state of the art -- including supervised models -- using no manually annotated data. We evaluate the method on several morphologically rich languages.

Originalspråk	engelska
Utgivningsår	2020
Status	Publicerad - 2020
MoE-publikationstyp	Ej behörig

Bibliografisk information

This paper contains corrigenda to a previously published paper (Hoya Quecedo et al., 2020). It corrects a mistake in the original evaluation setup, and the results reported in Section 6., in Tables 5, 6, and 7

Vetenskapsgrenar

113 Data- och informationsvetenskap

Åtkomst av dokument

https://arxiv.org/abs/2007.06104Licens: !!Unspecified

Revita: Language learning and AI
Yangarber, R., Katinskaia, A., Hou, J., Furlan, G. & Kylliäinen, I. P.
Projekt: Forskningsprojekt
LLL: Language Learning Lab
Yangarber, R., Katinskaia, A., Hou, J., Furlan, G. & Kylliäinen, I. P.
Projekt: Forskningsprojekt

Citera det här

@misc{bcd3ed0e5f274ad1a579da57a6cf7294,

title = "Neural disambiguation of lemma and part of speech in morphologically rich languages",

abstract = "We consider the problem of disambiguating the lemma and part of speech of ambiguous words in morphologically rich languages. We propose a method for disambiguating ambiguous words in context, using a large un-annotated corpus of text, and a morphological analyser -- with no manual disambiguation or data annotation. We assume that the morphological analyser produces multiple analyses for ambiguous words. The idea is to train recurrent neural networks on the output that the morphological analyser produces for unambiguous words. We present performance on POS and lemma disambiguation that reaches or surpasses the state of the art -- including supervised models -- using no manually annotated data. We evaluate the method on several morphologically rich languages.",

keywords = "113 Computer and information sciences, Language Technology",

author = "Quecedo, {Jos{\'e} Mar{\'i}a Hoya} and Koppatz, {Maximilian W.} and Giacomo Furlan and Roman Yangarber",

note = "This paper contains corrigenda to a previously published paper (Hoya Quecedo et al., 2020). It corrects a mistake in the original evaluation setup, and the results reported in Section 6., in Tables 5, 6, and 7",

year = "2020",

language = "English",

publisher = "arXiv.org",

address = "Unknown",

type = "Other",

}

TY - GEN

T1 - Neural disambiguation of lemma and part of speech in morphologically rich languages

AU - Quecedo, José María Hoya

AU - Koppatz, Maximilian W.

AU - Furlan, Giacomo

AU - Yangarber, Roman

N1 - This paper contains corrigenda to a previously published paper (Hoya Quecedo et al., 2020). It corrects a mistake in the original evaluation setup, and the results reported in Section 6., in Tables 5, 6, and 7

PY - 2020

Y1 - 2020

N2 - We consider the problem of disambiguating the lemma and part of speech of ambiguous words in morphologically rich languages. We propose a method for disambiguating ambiguous words in context, using a large un-annotated corpus of text, and a morphological analyser -- with no manual disambiguation or data annotation. We assume that the morphological analyser produces multiple analyses for ambiguous words. The idea is to train recurrent neural networks on the output that the morphological analyser produces for unambiguous words. We present performance on POS and lemma disambiguation that reaches or surpasses the state of the art -- including supervised models -- using no manually annotated data. We evaluate the method on several morphologically rich languages.

AB - We consider the problem of disambiguating the lemma and part of speech of ambiguous words in morphologically rich languages. We propose a method for disambiguating ambiguous words in context, using a large un-annotated corpus of text, and a morphological analyser -- with no manual disambiguation or data annotation. We assume that the morphological analyser produces multiple analyses for ambiguous words. The idea is to train recurrent neural networks on the output that the morphological analyser produces for unambiguous words. We present performance on POS and lemma disambiguation that reaches or surpasses the state of the art -- including supervised models -- using no manually annotated data. We evaluate the method on several morphologically rich languages.

KW - 113 Computer and information sciences

KW - Language Technology

M3 - Other contribution

PB - arXiv.org

ER -

Neural disambiguation of lemma and part of speech in morphologically rich languages

Sammanfattning

Bibliografisk information

Vetenskapsgrenar

Åtkomst av dokument

Projekt

Revita: Language learning and AI

LLL: Language Learning Lab

Citera det här