Translation services—Post-editing of machine translation output—Requirements
other information
drafter:Liu Zhiyang, Chai Ying, Luo Huifang, Zhang Jing, Ye Jian, Yu Limei, Meng Yongye, Zhang Chunliang, Zhang Xuetao, Zhu Xianchao, Zheng Chunping, Wang Haitao, He Zhongjun, Zhang Baolin, Huang Baorong, Zhu Li
Drafting unit:China National Institute of Standardization, Translators Association of China, Global Tone Communication Technology Co., Ltd., Shanghai Yizhe Information Technology Co., Ltd., Shanghai Youyi Information Technology Co., Ltd., Beijing Yueer Information
Focal point unit:National Technical Committee on Language and Terminology Standardization (SAC/TC 62)
Proposing unit:National Technical Committee on Language and Terminology Standardization (SAC/TC 62)
Publishing department:State Administration for Market Regulation National Standardization Administration
Some standard content:
ICS03.080.99;35.240.30
CCS A 10
National Standard of the People's Republic of China
GB/T40036—2021/IS018587:2017 Translation services-
Post-editing of machine translation output-Requirements(ISO 18587:2017.IDT)
Published on April 30, 2021
State Administration for Market Regulation
National Administration of Standardization
Implemented on November 1, 2021bZxz.net
GB/T 40036—2021/IS018587:2017 Pre
Normative references
3 Terms and definitions
Concepts related to machine translation
Concepts related to language and content
Concepts related to people or organization
Concepts related to translation
4 Post-editing process
Pre-production process
Production process
Goals of the post-editing process
Requirements for post-editing results
Tasks of post-editor
1.1 Post-production process
Final Verification and delivery
5 Competence and qualifications of post-editors
Professional quality
6 Requirements for in-depth post-editing
Appendix A (informative) Training of post-editors Appendix B (informative)
Light post-editing
Appendix C (informative)
Pre-editing
Requirements for the agreement and project signed between the client and the translation service provider Appendix D (informative)
Appendix E (informative)
References
Baidong post-editing
-rrKaeerKa-
GB/T 40036—2021/1SO18587:2017 This document is drafted in accordance with the provisions of GB/T1.12020% "Guidelines for standardization work - Part 1: Structure and drafting rules for standardization documents".
Post-editing requirements for machine translation results》This document uses the translation method equivalent to the IS018587:2017 translation service. Please note that some content of this document may involve patents. The issuing organization of this document does not assume the responsibility for identifying patents. This document was proposed and managed by the National Technical Committee for Language Production and Terminology Standardization (SAC/T (62)). The drafting units of this document are: China National Institute of Standardization, China Translators Association, China Global Tone Communication Technology Co., Ltd., Haiyijun Information Technology Co., Ltd., Shanghai Youyi Information Technology Co., Ltd., Beijing Yueer Information Technology Co., Ltd., Sichuan Language Bridge Information Technology Co., Ltd., Suzhou Lianyue Technology Co., Ltd., Shenyang Yayi Network Technology Co., Ltd., Beijing University of Posts and Telecommunications, Beijing Baidu Network Technology Co., Ltd., Shanghai Zhishanhe Network Technology Co., Ltd.,
The main drafters of this document are: Liu Zhiyang, Chai Ying, Luo Huifang, Zhang Jing, Ye Jian, Yu Limei, Meng Yongye, Zhang Chunliang, Zhang Xuetao, Zhu Xianchao, Zheng Chunping, Gong Haitao, He Zhongjun, Zhang Baolin, Huang Baorong, Zhu Li. -riKaeerkAca-||t t||GB/T40036—2021/IS018587:2017 Introduction
In order to meet the growing needs of the translation and localization industry, the use of machine translation systems has become more and more popular. Many translation service providers and clients have realized that when work time is very tight and/or budget is limited, the use of machine translation systems is a viable solution to complete translation projects: using machine translation systems, clients can obtain content that might not have been translated before, reduce translation costs, quickly launch products in specific markets, and speed up the flow of information. On the other hand, translation service providers can a) improve translation efficiency;
b) shorten delivery cycles:
Stay competitive in an environment where customers are increasingly demanding machine translation. However, there is no machine translation system that is exactly the same as human translation. translation system. Therefore, the final quality of the translation results still depends on the human translator and his/her post-editing ability.
Machine translation systems are evolving very quickly, and it is not realistic to develop a universal international standard for machine translation systems, as doing so may stifle innovation or be abandoned by the translation technology development industry. Therefore, this document only specifies the process for manual operations (i.e. post-editing) after the machine translation results are delivered. rrKaeerkAca-
1 Scope
Translation services
GB/T40036—2021/1S018587:2017 Post-editing of machine translation results
This document specifies the process requirements for in-depth post-editing of machine translation results and the ability requirements for post-editing personnel.
This document is intended for translators This document is intended for use by translation service providers (TSPs), clients and post-editors. This document applies only to content processed by a machine translation system. Note: For general translation services, see 1S017100. 2 Normative references This document has no normative references 3 Terms and definitions The following terms and definitions apply to this document IS( and IE are responsible for maintaining the terminology database used in the standard, which is accessed through the following two addresses: - IEC term database: Elcctropcdia: http://clcctropedia.org/ - ISs0 online browsing platform: http://iso.org/obp 3.1 Concepts related to machine translation machine translation; MT Machine translation The automatic translation of text (3.2.6) from one natural language into another natural language using a computer application (3.1, 2). 【Source: IS017100:2015, 2.2.2. Modified. Deleted the content of “speech translation”. Because it is not relevant to this document, “automated” was changed to “automatic” to avoid confusion with translation memory. 3.1.2
machine translation resultmachinetranslationoutput: MToutputThe result of machine translation (3.1.1).
LSource: ISO17100:2015, 2.2.3. Modified. — Changed outcome to “result”3.1.3
machine translation system
machine translation system
Technology used to implement machine translation (3.1.1) 3.1.4
post-editing
post-edit
Editing and stopping machine translation results (3.1.2) [Source: 1S017100:2015.2.2.1. Modified - Note deleted
-rKaeerkca-
GB/T40036—2021/IS018587:20173.1.5
full post-editingfull post-editing The process of post-editing (3.1.1) performed to obtain a translation that is equivalent to a human translation (3.1.3). 3.1.6
light post-editing
light post-editing
The process of post-editing (3.1.4) that does not require a translation product equivalent to that of a human translator (3.4.3) but only aims to obtain an understandable text.
3.2 Concepts related to language and content
Content
Information in any form
Examples: text, video or video.
Source languagesource language
The language of the content (3.2.1) to be translated (3.4.1). 3.2.3
Source language content
Fsource language content
The language (3.2.1) to be translated (3.4.1): [Source: 1S0[7100:2015,2.2.3
large langguagt
target language
the language into which the source language content (3.2.3) is translated (3.4.1).IS017100.2015.2.3.6
target language contenttarget language contenttext content (3.2.1) translated (3.1.1) from source language content (3.2.3), [Source: 1SO17100.2015, 2.3.3
texttext
content in written form (3.2.1)
[Source: 1S017100.2015.2.3.1
natural languagenatural language:NI
Source unknown, language that continues to develop in a specific way commonly used by humans to communicate. LSource: ISO/TS24620, 2.12
controlled natural language language: CNL controlled language controlled language
natural language that has been defined and restricted in its grammar and dictionary to reduce or eliminate ambiguity and complexity (3.2.7) Note 1: Controlled natural language is a general term, an uncountable noun, referring to the abstract properties of all controlled natural speech, rather than a particular natural speech or the basis for a specific process. Controlled natural language is a language that has been processed (i.e. constructed) to reduce or eliminate ambiguity and complexity, to help human readers (especially non-language users, non-experts and people with limited understanding) read text (3.2.6), and to make text more suitable for computer processing. 2
-rrK aeerkAca-
GB/T40036—2021/IS018587:2017 Note 2: A controlled natural language is a language that has been processed (i.e. constructed) after a specific natural language has been processed: it is more restricted in terms of vocabulary, syntax or meaning, but at the same time retains most of its natural properties. In this case, a controlled natural language can be a number of different names, [Source: ISO/TS24620.2.6]
segment
a unit of text (3.2.6) generated by a computer application for translation Note: A segment can be a sentence title or other text unit, such as a phrase, a single word or a single character. 3.2.10
Regional characteristics
lo cale
The set of characteristics, information or customs of the target audience in terms of language, culture, technology and regional customs. 3.2.11
languageregister
The type of language used on a specific day or in a specific social or industry field. 3.3 Concepts related to people or organizations
clientclient
Customercustomer
The person or organization who entrusts the translation service provider (3.3.5) with the service through a formal agreement. Note: The client may be an individual or organization seeking or purchasing services and may be external or internal to the translation service provider (3.3.5). [Source: 1S01 7100:2015,2.4.3]3.3.2
Translator
A person who translates (3.4.1).
[Source: 1S017100:2015,2.4.1
Reviser
A person who reviews the translation results (3.4.4).
[Source: 1S017100:2015,2.4.5. Modified—Since it has been explained in the definition of "review", it is not necessary to specify here"
Post-editor
A person who performs post-editing (3.1.4). 3.3.5
translation service provider: TSP
A service provider that provides translation services (3.4.5). Note: A translation service provider may be a translation company, translation agency, translation organization (for-profit, pre-profit or government), in-house translator or post-editor, or an in-house translation department.
LSource: Definitions cited in IS017100: 2015.2.4.2, with modifications" 3.3.6
projectmanager;PM
Project manager
The person responsible for managing the relevant aspects of a translation project or a post-editing project and its processes [Source: IS017100: 2015.2.4.8. With modifications - "post-editing" was added to be more relevant to the content of this document 3
-rrKaeerkAca-
GB/T40036—2021/IS018587:20173.4 Concepts related to translation
Translation Translate
Convert source language content (3.2.3) into target language content (3.2.5) in written form. L Source: ISO17100:2015.2.1.1.
Translation (translation process)
translation
A set of processes for converting source language content (3.2.3) into target language content (3.2.5) [Source: IS017100:2015, 2.1.2. Modified—no need to mention *1 and “form”: In addition, to avoid confusion with the definition of translation result, the note has been deleted.
Human translation
Translation (3.4.2) performed by a translator (3.3.2) 3.4.4
translation output
translation result
the result of translation (3.4.2).
translation service
translation service
the intangible product resulting from the interaction between the client (3.3.1) and the translation service provider (3.3.5): [Source: IS017100:20[5,2.1.6]3.4.6
translation memory; TM
translation memory
a database of matching segments (3.2.9) in the source language (3.2.2) and the target language (3.2.4). NOTE 1 The purpose of a translation memory is to facilitate the retrieval and use of previously translated material. 3.4.7
revision
Perform a bilingual check of the target language content (3.2.5) and the source language content (3.2.3) to determine their suitability for the agreed purpose. Note: "Bilingual editing" is sometimes used as a synonym for bilingual proofreading. [Source: IS7100.2015.2.2.6
Verification
Confirmation by the project manager (3.3.6) of the translation service provider (3.3.5) whether the requirements of the project are met. Source: IS017100:20152.5.1
4 Post-editing process
4.1 General
"Translation service provider" has been added for clarity. Post-editing is done on the basis of machine translation results, aiming to check the accuracy and comprehensibility of machine translation, improve the text, improve the readability of the text, and eliminate errors. Post-editing is different from translation because it involves three texts: source text, machine translation results and final standard text
There are two levels of post-editing: light post-editing and deep post-editing: which level to use Depending on the purpose of the translation results and the requirements of the client, this document specifies deep post-editing, while Appendix B specifies light post-editing. 4
-rrKaeerkAca-
GB/T40036—2021/1S018587:2017 The post-editing process can be automatic post-editing or manual post-editing. Appendix E explains automatic post-editing. Manual translation editing refers to the process of manual editing of machine translation results by post-editors. The post-editing process can occur in an environment where all the texts to be processed have been processed by the machine translation system, or in a multi-mode translation work environment that integrates machine translation, translation memory and terminology management. Whether the results of post-editing are entered into the translation memory (automatic entry or partial entry) depends on the translation service provider's process and its client's quality assurance requirements.
4.2 Pre-production process
The translation service provider shall have relevant processes and, when necessary, consult with the client to determine whether the source language content is suitable for machine translation and subsequent post-editing, because the combined efficiency of machine translation/post-editing depends on the machine translation system, language combination, field and style of the source language content:
The source language content may be pre-edited before machine translation to facilitate machine processing and improve the quality of the original machine translation results, thereby reducing the workload of post-editing. Appendix (Explanation of pre-translation editing) The translation service provider and the client reach an agreement and keep records. If the agreement is reached verbally or by telephone, the translation service provider shall confirm the agreement and terms in a formal form (for example, by letter, fax or email). Whether or not it is contractual, the agreement shall include or reference commercial terms and requirements. The agreement may also be required to be consistent with the relevant content of this document. Appendix 1) lists the list of elements that may be included in the agreement.
Before taking any action that deviates from the original agreement, all parties shall agree on such action and the agreed amendments shall be recorded and retained with the original agreement document. The TSP shall ensure that the requirements for the post-editing project are identified, recorded and made available to the post-editor: The TSP shall also ensure that the post-editor is aware of the relevant requirements of the project (e.g. quality level and target audience, etc.) and appropriate examples, and complies with them.
Note 1: The requirements for the post-editing project may be the requirements stated by the client or part of the TSP's service procedures, or a combination of both. The TSP shall ensure that the source content and the text to be post-edited are in a format that can be used by the post-editor. Typically, this means files generated in standard software applications for editing the text (in the case of editing the original text) or for reading the text (in the case of reference to the source content). The TSP, in consultation with the client, shall ensure that the post-editor has access to all reference materials and any resources to perform the post-editing.
The translation service provider shall inform the post-editor of the expected availability of machine translation results: The translation service provider shall ensure that the text to be post-edited is marked with relevant instructions/markers to accurately distinguish between machine translation results and translation results from other sources (e.g., translation memory): Note 2: The text to be post-edited can be the output of a machine translation system or the translation result that integrates different sources (e.g., translation memory and machine translation).
4.3 Production process
4.3.1 Objectives of the post-editing process
The translation service provider shall ensure that the post-editing staff always meets the following objectives during the post-editing process: a) Interpretability of the post-editing results; b) Symmetry between the source language content and the target language content; c) Compliance with the post-editing requirements and project requirements established by the translation service provider. 4.3.2 Requirements for post-editing results
The translation service provider shall ensure that the following requirements are met: 5
-riKacerKAca-
GB/T40036—2021/IS()18587:2017a) Technical information/vocabulary- a) be consistent with, and adapt to, the terminology of the specific field: use standard grammar, spelling, punctuation, diacritics, special symbols and abbreviations, and conform to other spelling conventions of the target language; b) comply with applicable standards; d) be correctly formatted; c) be suitable for the target audience and the purpose of the spoken content: ) comply with the agreement signed between the client and the translation service provider. In addition, if the agreement signed between the client and the translation service provider and the project requirements include any of the following, the translation service provider will ensure that these requirements are met:
conform to the terminology and/or any other reference material provided by the client: - comply with any proprietary and/or client style guides (including register and regional characteristics); - comply with post-editing guidelines.
Note: For the agreement between the client and the translation service provider, see Appendix D4.3.3 Post-Editor's Tasks
The translation service provider shall ensure that the post-editor performs the following tasks: a) Read the machine translation results and evaluate whether the original content needs to be rephrased; a) Use the source language content as a reference to interpret and correct the original content if necessary; e) Use existing elements from the machine translation results or provide new translations to generate the original content. 4.4 Post-Production Process
4.4.1 Final Verification and Delivery
The translation service provider shall have a process to verify the final product according to the project requirements before delivering the post-edited results to the client.
4.4.2 Feedback
The translation service provider shall have a process to obtain feedback from the post-editor on the performance of the machine translation system in order to improve the machine translation system and subsequent processes.
5 Competencies and qualifications of post-editors
The translation service provider shall ensure that the translation editors have the following competencies and qualifications. 5.1 Competencies
a) Translation competence: the ability to translate content, including the ability to deal with problems in the process of language content understanding and generation: and the ability to deliver the target language content in accordance with the agreement signed between the client and the translation service provider, the agreement signed between the translation service provider and the post-editor and other project requirements.
b) Linguistic and text processing skills in the source and target languages: the ability to understand the source language, use the target language proficiently, and master general or professional knowledge of the text type (if applicable), including the ability to use this knowledge to generate text or other target language content.
c) Research, acquisition and processing of information: the ability to efficiently develop the necessary language knowledge and expertise to better understand the source language content and translate it into the target language. Research competence also requires experience in using research tools and the ability to develop appropriate strategies to effectively use existing information resources. 6
rKaeerkAca-
GB/T40036—2021/1S018587:2017d) Cultural competence: The ability to possess knowledge of the source language and to apply relevant information such as behavioral standards, valuation systems, and regional characteristics that are consistent with the cultural characteristics of the source language and the target language.
c) Technical competence: The knowledge, abilities, and skills to utilize technical resources: including the use of tools and information technology (IT) systems to support the entire translation process, to complete the technical tasks in the translation process. f) Domain competence: The ability to understand and grasp the professional domain content generated in the source language and present it in the target language. 5.2 Qualifications
The translation service provider shall determine whether the post-translation editor meets the service qualifications of this document through the supporting documents provided by the post-translation editor. The post-translation editor shall meet at least one of the following criteria: a) a degree in translation, linguistics or language awarded by a recognized higher education institution, or an equivalent professional degree with sufficient translation training;
b) other professional degrees awarded by a recognized higher education institution and have the equivalent of two years of full-time professional translation or post-editing experience;
) have the equivalent of five years of full-time professional translation or post-editing experience. 5.3 Professional qualities
Post-editors should adopt a professional approach to post-editing tasks and: a) have general knowledge of machine translation technology and a basic understanding of common errors in machine translation systems; b) have general knowledge of computer-assisted translation tools; c) have the knowledge and ability to judge whether a project is feasible based on the time and effort required for post-editing; d) be able to follow instructions, focus on specific issues and make adjustments based on the content given. Post-editors should have the ability to provide structured feedback on frequent errors in machine translation results to promote the improvement of machine translation systems.
Post-editors should have a basic understanding of the relationship between terminology management systems and machine translation systems. Note: For content that can be included in post-editing personnel training, see Appendix A. 6 Requirements for in-depth post-editing
For in-depth post-editing, the results should be accurate, understandable, and appropriate in style, with correct syntax, notation, and punctuation. The goal is to produce a translation that is equivalent to the results of a human translation. Nevertheless, it is still recommended that post-editors use the results of machine translation as much as possible. For this level of post-editing, post-editors should focus on: a) ensuring that no information is added or omitted: modify any inappropriate content; restructure the text if the meaning is ambiguous or unclear: generate grammatically, syntactically and semantically correct standard language content; d) follow customer and/or domain-specific terminology specifications: follow spelling, punctuation and hyphenation rules; ensure that a style appropriate to the type of text is used and adhere to the style guide provided by the customer: g) h) follow formatting rules, -riKacerKAca-
GB/T40036—2021/IS0)18587:2017 Appendix A
(Informative)
Post-Editor Training
Training language professionals in the knowledge and skills required for post-editing can: a) help meet the growing demand for translation and shorten delivery cycles; provide language professionals with the skills required for post-editing, which are different from translation skills; h)
Improve language professionals' understanding of machine translation and its capabilities so that they can better adapt to post-editing projects in a machine translation environment;
d Promote the development and innovation of machine translation technology production Post-editor training may include:
Advanced use of translation memory and machine translation technology to enable the use of translation memory Post-editing based on the combined results of the terminology database and the machine translation system; training should include information on typical machine translation errors, such as style issues, word-for-word translation, grammatical errors (including negative translation errors, verb translation errors, etc.), translating names that should not be translated, etc.; in-depth terminology work, including terminology database management, for example, knowledge of various terminology management systems and terminology exchange formats such as TBX (TermBase eXchange):
Advanced text processing skills to be able to use macros and search and replace functions; practice of light post-editing and deep post-editing: use quality tools to perform quality checks at the end of the project: 8
-rKaeerkca-
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.