Multiword units in machine translation and translation technology

Monti, Johanna; Seretan, Violeta; Pastor, Gloria Corpas; Mitkov, Ruslan

doi:10.1075/cilt.341.01mon

Cited by 20 publications

(9 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Before considering taking advantage of these observations to try to improve NMT systems, a series of complementary analyses must be conducted. Indeed, this study has many limitations, such as focusing only on a subcategory of MWUs [20], on a single language pair, and on a single genre of texts. Moreover, a thorough qualitative analysis is essential to better understand the results and evaluate the proposed explanations.…”

Section: Discussionmentioning

confidence: 99%

“…Neural machine translation (NMT) systems are currently considered to bridge the gap between human and machine translation [22,26]. However, little research has been done to determine whether NMT systems are also very effective in processing multiword units [20,27], whereas the importance of preformed units in language use is now well established, including in foreign language learning and translation [1,21,24]. The present study addresses this issue by comparing formulaic language in human and neural machine translation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using CollGram to Compare Formulaic Language in Human and Neural Machine Translation

Bestgen

2021

Preprint

View full text Add to dashboard Cite

A comparison of formulaic sequences in human and neural machine translation of quality newspaper articles shows that neural machine translations contain less lower-frequency, but strongly-associated formulaic sequences, and more high-frequency formulaic sequences. These differences were statistically significant and the effect sizes were almost always medium or large. These observations can be related to the differences between second language learners of various levels and between translated and untranslated texts. The comparison between the neural machine translation systems indicates that some systems produce more formulaic sequences of both types than other systems.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Using CollGram to Compare Formulaic Language in Human and Neural Machine Translation

Bestgen

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Se utilizan para ello diversas pruebas o criterios como, por ejemplo, la imposibilidad de sustituci6n paradigmatica (*dar gato por conejo) o la existencia de equivalentes de traducci6n (dar gato por liebre "' EN pull the wool of someone's eyes) (cf. Ramisch 2015;Monti et al 2018), aunque el criterio mas empleado es el de frecuen cia (de aparici6n y/o de co-aparici6n).…”

Section: Detection Automaticaunclassified

25 El hablar y el discurso repetido: la fraseología

Mellado¹,

Pastor²,

Berty³

2021

Manual De Lingüística Del Hablar

View full text Add to dashboard Cite

El hablar y el discurso repetido: la fraseologiaResumen: Este capitulo muestra la interrelacion entre fijacion y variabilidad en las unidades fraseologicas desde distintos puntos de vista. En primer lugar, realizamos un analisis detallado del concepto de «discurso repetido» de Coseriu, que ya conside ra en su origen la idea de cambio creativo, para despues ofrecer una panoramica de la evolucion de la fraseologia en relacion a la lingilistica textual. En segundo lugar, se presenta una clasificacion de la tipologia de la variacion fraseologica, ilustrada con ejemplos de corpus lingiiisticos y centrada en los niveles del sistema y habla, asi como en la intencionalidad del hablante. En tercer lugar, tratamos el tema de la variabilidad fraseologica y el giro que ha tornado la nocion de «fijacion» desde que se dispone de datos masivos de corpus. En este contexto, las magnitudes de frecuencia absoluta, normalizada y de significacion estadistica desempeiian un papel f u ndamental para el grado de fijacion.Palabras clave: unidad fraseologica, desautomatizacion, corpus, variabilidad fraseo logica, procesamiento de lenguaje natural 1 Fraseologia y discurso repetido en la lingilistica del hablar 1.1 Lingllistica de texto integral, tecnica libre del discurso y discurso repetido La «lingilistica del sentido» de Coseriu se coneibe como una lingilistica de texto holis tica integrada en una «lingiiistica del hablar», que abarca las distintas disciplinas lin gilisticas (cf. Loureda Lamas 2007; 2009) y que ya fue esbozada por Coseriu en los aiios 1950 (Determinaci6n y entomo, 1955; cf. Coseriu 2007, 88). La lingiiistica del sen tido es <«tres veces lingiiistica del texto> (711 El hablar y las lenguas). Los textos son hechos individuales, pero no absolutamente particulares, pues presentan, ademas, una dimension universal, que incluye aquellos rasgos de la textualidad, y una dimen sion hist6rica que hace que compartan tradiciones y moldes expresivos» (Loureda La mas 2009, 73, con referenda a Coseriu/Loureda Lamas 2006). Justamente en esta ulti ma dimension tradicional del nivel del texto se localiza el fenomeno del «discurso repetido», junto a los tipos de texto y las tradiciones discursivas (cf. Loureda Lamas 2009, 83ss.). En el marco de la lingilistica de texto de Coseriu, el concepto del discurso repetido se asocia con las relaciones de «sentido» surgidas entre un texto o un seg-

show abstract

“…The correct interpretation of Multiword Expressions (MWEs) is crucial to many natural language processing (NLP) applications but is challenging and complex. In recent years, the computational treatment of MWEs has received considerable attention, but there is much more to be done before one can claim that NLP and Machine Translation (MT) systems process MWEs successfully [12].…”

Section: Introductionmentioning

confidence: 99%

Transformer-based Detection of Multiword Expressions in Flower and Plant Names

Damith¹,

Haddad²,

Ranasinghe³

et al. 2022

Proceedings of the International Conference EUROPHRAS 2022 (Short Papers, Posters and MUMTTT Workshop Contributions)

View full text Add to dashboard Cite

Multiword expression (MWE) is a sequence of words which collectively present a meaning which is not derived from its individual words. MWEs detection is an important topic in different natural language processing (NLP) applications, including machine translation. Therefore, detecting MWEs in different domains is an important research topic. With this study, we explore state-of-the-art neural transformers in the task of detecting MWEs in flower and plant names. We evaluate different transformer models on a dataset created from Encyclopedia of Plants and Flower. We empirically show that neural transformers could outperform previous neural models like long-short term memory (LSTM) even in specific domains such as flower and plant names.

show abstract

Multiword units in machine translation and translation technology

Cited by 20 publications

References 55 publications

Using CollGram to Compare Formulaic Language in Human and Neural Machine Translation

Using CollGram to Compare Formulaic Language in Human and Neural Machine Translation

25 El hablar y el discurso repetido: la fraseología

Transformer-based Detection of Multiword Expressions in Flower and Plant Names

Contact Info

Product

Resources

About