The efficacy of human post-editing for language translation

@article{Green2013TheEO,
  title={The efficacy of human post-editing for language translation},
  author={Spence Green and Jeffrey Heer and Christopher D. Manning},
  journal={Proceedings of the SIGCHI Conference on Human Factors in Computing Systems},
  year={2013}
}
Language translation is slow and expensive, so various forms of machine assistance have been devised. Automatic machine translation systems process text quickly and cheaply, but with quality far below that of skilled human translators. To bridge this quality gap, the translation industry has investigated post-editing, or the manual correction of machine output. We present the first rigorous, controlled analysis of post-editing and find that post-editing leads to reduced time and, surprisingly… 
View on ACM
idl.cs.washington.edu
179 Citations
Highly Influential Citations
19
Background Citations
99
Methods Citations
30
Results Citations
21

179 Citations

Citation Type

Citation Type

Predictive translation memory: a mixed-initiative system for human language translation
TLDR
In a large-scale study, it is found that professional translators are slightly slower in the interactive mode yet produce slightly higher quality translations despite significant prior experience with the baseline post-editing condition.
Translation Quality and Effort: Options versus Post-editing
TLDR
This paper directly compares two common assistance types – selection from lists of translation options, and postediting of machine translation output produced by Google Translate – across two significantly different subject domains for Chinese-to-English translation.
Analysis of Edit Operations for Post-editing Systems
TLDR
A novel approach for seeking suitable pairs of n-grams when recommending n- Grams based on the type of text (manual or administrative) and MT system used for machine translation is introduced.
Multi-Modal Approaches for Post-Editing Machine Translation
TLDR
It is argued that future translation environment research should focus more strongly on pen, touch, and speech modalities in addition to mouse- and keyboard-based approaches, as the task changes from mainly generating text to correcting errors within otherwise helpful translation proposals.
Human Effort and Machine Learnability in Computer Aided Translation
TLDR
Evaluation with professional bilingual translators shows that post-edit is faster than interactive at the cost of translation quality for French-English and EnglishGerman, and analysis shows that tuning directly to HTER results in fine-grained corrections to subsequent machine output.
Translators’ perceptions of literary post-editing using statistical and neural machine translation
In the context of recent improvements in the quality of machine translation (MT) output and new use cases being found for that output, this article reports on an experiment using statistical and…
Effects of word alignment visualization on post-editing quality & speed
TLDR
Human adequacy judgements of the post-edited sentences are obtained, and it is demonstrated that when machine translation quality is low, post-editing quality is consistently higher, by a statistically significant amount, when bilingual post-edsitors are presented with alignment data.
Productivity and quality when editing machine translation and translation memory outputs: an empirical analysis of English to Welsh translation
TLDR
Findings in a controlled study carried out to examine the possible benefits of editing Machine Translation and Translation Memory outputs when translating from English to Welsh contradict supposed similarities between translation quality in terms of style and post-editing Machine Translation.
Post-editing Effort of a Novel With Statistical and Neural Machine Translation
TLDR
This first experiment in the literature in which a novel is translated automatically and then post-edited by professional literary translators is conducted, and both MT approaches result in increases in translation productivity: PBMT by 18%, and NMT by 36%.
How Does the Post-Editing of Neural Machine Translation Compare with From-Scratch Translation?: A Product and Process Study
TLDR
The three main findings were: 1) post-editing GNMT was only significantly faster than fromscratch translation for domain-specific texts, but it significantly reduced the participants’ cognitive effort for both text types; and 2) the student translators generally showed a positive attitude towards post-edsiting, but they also pointed to various challenges in the post-EDiting process.
...
...

References

SHOWING 1-10 OF 61 REFERENCES
A process study of computer-aided translation
TLDR
The computer-aided tool Caitra is developed that makes suggestions for sentence completion, shows word and phrase translation options, and allows postediting of machine translation output.
Translating by post-editing: is it the way forward?
  • I. Garcia
  • Computer Science
    Machine Translation
  • 2011
TLDR
It is discussed whether translators should consider post-editing as a viable alternative to conventional translation, as it produces significantly better statistical results compared to translating manually.
Repairing Texts: Empirical Investigations of Machine Translation Post-Editing Processes
TLDR
In Repairing Texts, Hans P. Krings challenges the idea that, given the effectiveness of machine translation, major costs could be reduced by using monolingual staff to post -- edit translations.
Post-editing machine translated text in a commercial setting: Observation and statistical analysis
TLDR
A mixed method approach was employed to both quantitatively and qualitatively analyse the data and gain detailed insights into the post-editing activity from various view points, indicating that a number of factors, such as sentence structure, document component types, use of product specific terms, and post-edsiting patterns and behaviour, have effect on the amount of post-EDiting effort in an intertwined manner.
Pauses as Indicators of Cognitive Effort in Post-editing Machine Translation Output
TLDR
This paper presents data from a research project which includes an analysis of pauses in post-editing, triangulated with the Choice Network Analysis method and Translog, and suggests that while pauses provide some indication of cognitive processing, supplementary methods are required to give a fuller picture.
A Productivity Test of Statistical Machine Translation Post-Editing in a Typical Localisation Context
TLDR
A Productivity Test of Statistical Machine Translation Post-Editing in a Typical Localisation Context and results show a productivity increase for each participant, with significant variance across inviduals.
Enabling Monolingual Translators: Post-Editing vs. Options
TLDR
This study carried out a study on monolingual translators with no knowledge of the source language, but aided by post-editing and the display of translation options, able to translate 35% of Arabic and 28% of Chinese sentences correctly on average.
Collaborative translation by monolinguals with machine translators
In this paper, we present the concept for collaborative translation, where two non-bilingual people who use different languages collaborate to perform the task of translation using machine…
Interactive translation vs pre-translation in the context of translation memory systems: Investigating the effects of translation method on productivity, quality and translator satisfaction
TLDR
As the demand for translation continues to rise, more and more translators are looking to TMs to help increase productivity; however, for a variety of reasons, such as cost and incompatible file formats, they do not always have access to a useful TM.
Long Distance Revisions in Drafting and Post-editing
TLDR
Property of translation processes, as observed in the translation behaviour of student and professional translators, are investigated and it is suggested how those findings might be taken into account in the design of computer assisted translation tools.
...
...