This repository contains 46 examples of English source documents (SDs) in various content domains that are often dealt with by translation service providers (TSPs) and several types of translation-related derivatives that we have annotated and produced aiming to analyze the norms and competences in the translation process.
Textual data of source documents are found here.
For all of them, we have so far produced the following types of derivatives (See README.md in each directory for the details).
- Professional translation
- Post-edited machine translation outputs
- Machine translation outputs generated by TexTra on December 2020
- Post-edited version of the above machine translation output
- Automatically derived paragraph-level HTER scores and HTER-based word-level binary labels
- Span-based issue annotations based on MQM-like manual quality assessment of MT
- Unpolished/Polished translation pairs
Figure 1 summarizes the processes of producing these heterogeneous translations and what can be analyzed by comparing some of them.
- MT outputs vs. their post-edited version: the issues of MT
- Human translation (HT) vs. MT+PE: the gap between what can be assured by ISO/TC37 (2015) and ISO/TC37 (2017)
- Draft/unpolished translation vs. final/polished translation: the art in human translation (or translation strategies)
Fig. 1: Overview of the processes of producing translations and their purposes.
- Include other types of derivatives.
- Extend the number of documents as well as translation directions.
- ISO/TC37. ISO 17100:2015 Translation Services: Requirements for Translation Services. 2015.
- ISO/TC37. ISO 18587:2017 Translation Services: Post-editing of Machine Translation Output: Requirements. 2017.
- We collected the source documents manually. See this JSON file for their origin and license.
- Use and/or redistribution of this dataset, except for the source documents, is permitted under the conditions of Creative Commons Attribution-NonCommercial-ShareAlike License 4.0.
- Please carefully read README.md in subdirectories and precisely confirm the attribution (BY) of each outcome.
This dataset is an outcome of JSPS KAKENHI Grant-in-Aid for Scientific Research (S) 19H05660, Developing a Translation Process Model and Constructing an Integrated Translation Environment through Detailed Descriptions of Translation Norms and Competences.