Using Error Annotation to Support Post-Editing in the Context of Translator Training

Justus Brockmann; Claudia Karin Wiesinger; Alina Secara; Dragos Ioan Ciobanu

Using Error Annotation to Support Post-Editing in the Context of Translator Training

Department for Translation Studies

Publications: Contribution to book › Contribution to proceedings › Peer Reviewed

Abstract

Translation quality assessment (TQA), which frequently involves error annotation
based on a standardised typology, is a relevant skill for professional translators.
However, previous research has identified a TQA competence gap between translation-related academic programmes and industry needs. We analyse annotation and questionnaire data from an experiment conducted with 17 students performing an error annotation and post-editing task on English to German neural machine translation output. Our data show that, similar to a Gold Standard, the top five errors annotated by students were: Mistranslation, Other, Unidiomatic, Grammatical Register, and Grammar. Nevertheless, students used the Other category more frequently, and occasionally incorrectly. Moreover, the variety of error
types used by students overall was
considerably higher than in the Gold
Standard, which suggests a certain level of
insecurity about the annotation task.
Although this phenomenon is also
mentioned in the questionnaire responses,
most students did explicitly perceive error
annotation to be a useful task to support
post-editing.

Original language	English
Title of host publication	Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022
Subtitle of host publication	4-6 July 2022, Rhodes Island, Greece.
Editors	Sheila Castilho, Rocío Caro Quintana, Maria Stasimioti, Vilelmini Sosoni
Place of Publication	Schumen
Publisher	INCOMA Ltd.
Pages	113-124
Publication status	Published - 2023
Event	New Trends in Translation & Technology - Sheraton Rhodes Resort, Rhodes, Greece Duration: 4 Jul 2022 → 6 Jul 2022 https://nettt-conference.com/

Publication series

Series	Proceedings of the Conference New Trends in Translation and Technology.
ISSN	2815-4711

Conference

Conference	New Trends in Translation & Technology
Abbreviated title	NeTTT
Country/Territory	Greece
City	Rhodes
Period	4/07/22 → 6/07/22
Internet address	https://nettt-conference.com/

Austrian Fields of Science 2012

602051 Translation studies

Access to Document

https://acl-bg.org/proceedings/2022/NeTTT%202022/NeTTT-2022-Final-Proceedings.pdfLicence: CC BY-NC-ND 4.0

Using Error Annotation to Support Post-Editing in the Context of Translator Training
Brockmann, J. (Speaker), Wiesinger, C. K. (Contributor), Secara, A. (Contributor) & Ciobanu, D. I. (Contributor)
4 Jul 2022
Activity: Talks and presentations › Talk or oral contribution › Science to Science

Cite this

Brockmann, J., Wiesinger, C. K., Secara, A., & Ciobanu, D. I. (2023). Using Error Annotation to Support Post-Editing in the Context of Translator Training. In S. Castilho, R. Caro Quintana, M. Stasimioti, & V. Sosoni (Eds.), Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022: 4-6 July 2022, Rhodes Island, Greece. (pp. 113-124). INCOMA Ltd.. https://acl-bg.org/proceedings/2022/NeTTT%202022/NeTTT-2022-Final-Proceedings.pdf

Brockmann, Justus ; Wiesinger, Claudia Karin ; Secara, Alina et al. / Using Error Annotation to Support Post-Editing in the Context of Translator Training. Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022: 4-6 July 2022, Rhodes Island, Greece.. editor / Sheila Castilho ; Rocío Caro Quintana ; Maria Stasimioti ; Vilelmini Sosoni. Schumen : INCOMA Ltd., 2023. pp. 113-124 (Proceedings of the Conference New Trends in Translation and Technology.).

@inproceedings{4a2be95b4cd34c21a5576dcb875b3b29,

title = "Using Error Annotation to Support Post-Editing in the Context of Translator Training",

abstract = "Translation quality assessment (TQA), which frequently involves error annotation based on a standardised typology, is a relevant skill for professional translators. However, previous research has identified a TQA competence gap between translation-related academic programmes and industry needs. We analyse annotation and questionnaire data from an experiment conducted with 17 students performing an error annotation and post-editing task on English to German neural machine translation output. Our data show that, similar to a Gold Standard, the top five errors annotated by students were: Mistranslation, Other, Unidiomatic, Grammatical Register, and Grammar. Nevertheless, students used the Other category more frequently, and occasionally incorrectly. Moreover, the variety of error types used by students overall was considerably higher than in the Gold Standard, which suggests a certain level of insecurity about the annotation task. Although this phenomenon is also mentioned in the questionnaire responses, most students did explicitly perceive error annotation to be a useful task to support post-editing.",

author = "Justus Brockmann and Wiesinger, \{Claudia Karin\} and Alina Secara and Ciobanu, \{Dragos Ioan\}",

year = "2023",

language = "English",

series = "Proceedings of the Conference New Trends in Translation and Technology.",

publisher = "INCOMA Ltd.",

pages = "113--124",

editor = "Sheila Castilho and \{Caro Quintana\}, Roc{\'i}o and Maria Stasimioti and Vilelmini Sosoni",

booktitle = "Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022",

note = "New Trends in Translation \& Technology, NeTTT ; Conference date: 04-07-2022 Through 06-07-2022",

url = "https://nettt-conference.com/",

}

Brockmann, J, Wiesinger, CK , Secara, A & Ciobanu, DI 2023, Using Error Annotation to Support Post-Editing in the Context of Translator Training. in S Castilho, R Caro Quintana, M Stasimioti & V Sosoni (eds), Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022: 4-6 July 2022, Rhodes Island, Greece.. INCOMA Ltd., Schumen, Proceedings of the Conference New Trends in Translation and Technology., pp. 113-124, New Trends in Translation & Technology, Rhodes, Greece, 4/07/22. <https://acl-bg.org/proceedings/2022/NeTTT%202022/NeTTT-2022-Final-Proceedings.pdf>

Using Error Annotation to Support Post-Editing in the Context of Translator Training. / Brockmann, Justus; Wiesinger, Claudia Karin ; Secara, Alina et al.
Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022: 4-6 July 2022, Rhodes Island, Greece.. ed. / Sheila Castilho; Rocío Caro Quintana; Maria Stasimioti; Vilelmini Sosoni. Schumen: INCOMA Ltd., 2023. p. 113-124 (Proceedings of the Conference New Trends in Translation and Technology.).

Publications: Contribution to book › Contribution to proceedings › Peer Reviewed

TY - GEN

T1 - Using Error Annotation to Support Post-Editing in the Context of Translator Training

AU - Brockmann, Justus

AU - Wiesinger, Claudia Karin

AU - Secara, Alina

AU - Ciobanu, Dragos Ioan

PY - 2023

Y1 - 2023

N2 - Translation quality assessment (TQA), which frequently involves error annotation based on a standardised typology, is a relevant skill for professional translators. However, previous research has identified a TQA competence gap between translation-related academic programmes and industry needs. We analyse annotation and questionnaire data from an experiment conducted with 17 students performing an error annotation and post-editing task on English to German neural machine translation output. Our data show that, similar to a Gold Standard, the top five errors annotated by students were: Mistranslation, Other, Unidiomatic, Grammatical Register, and Grammar. Nevertheless, students used the Other category more frequently, and occasionally incorrectly. Moreover, the variety of error types used by students overall was considerably higher than in the Gold Standard, which suggests a certain level of insecurity about the annotation task. Although this phenomenon is also mentioned in the questionnaire responses, most students did explicitly perceive error annotation to be a useful task to support post-editing.

AB - Translation quality assessment (TQA), which frequently involves error annotation based on a standardised typology, is a relevant skill for professional translators. However, previous research has identified a TQA competence gap between translation-related academic programmes and industry needs. We analyse annotation and questionnaire data from an experiment conducted with 17 students performing an error annotation and post-editing task on English to German neural machine translation output. Our data show that, similar to a Gold Standard, the top five errors annotated by students were: Mistranslation, Other, Unidiomatic, Grammatical Register, and Grammar. Nevertheless, students used the Other category more frequently, and occasionally incorrectly. Moreover, the variety of error types used by students overall was considerably higher than in the Gold Standard, which suggests a certain level of insecurity about the annotation task. Although this phenomenon is also mentioned in the questionnaire responses, most students did explicitly perceive error annotation to be a useful task to support post-editing.

M3 - Contribution to proceedings

T3 - Proceedings of the Conference New Trends in Translation and Technology.

SP - 113

EP - 124

BT - Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022

A2 - Castilho, Sheila

A2 - Caro Quintana, Rocío

A2 - Stasimioti, Maria

A2 - Sosoni, Vilelmini

PB - INCOMA Ltd.

CY - Schumen

T2 - New Trends in Translation & Technology

Y2 - 4 July 2022 through 6 July 2022

ER -

Brockmann J, Wiesinger CK , Secara A , Ciobanu DI. Using Error Annotation to Support Post-Editing in the Context of Translator Training. In Castilho S, Caro Quintana R, Stasimioti M, Sosoni V, editors, Proceedings of the New Trends in Translation and Technology Conference - NeTTT 2022: 4-6 July 2022, Rhodes Island, Greece.. Schumen: INCOMA Ltd. 2023. p. 113-124. (Proceedings of the Conference New Trends in Translation and Technology.).

Using Error Annotation to Support Post-Editing in the Context of Translator Training

Abstract

Publication series

Conference

Austrian Fields of Science 2012

Access to Document

Fingerprint

Activities

Using Error Annotation to Support Post-Editing in the Context of Translator Training

Cite this