Post-Editing Machine Translation
Post-editing skills have thrived as an asset for translators in the language industry. Machine Translation (MT) has considerably improved in popularity, as the demand for post-editing services has also grown in line with this. TechNavio even predicted that the MT market would attribute a rapid increasing content volume. The Growth of Post-Editing Machine Translation
Post-Editing Machine Translation (PEMT) facilitates the process of ‘fixing’ machine translation outputs to bring it closer to the standard of human translation. The development of this tool manages to assist post-editors avoiding exhaustive attempts in producing standard translation products. Frankly speaking, translation industry cannot only rely on the machine translation as the end of its technological adaptation. Beyond translation itself, post-editing is a crucial matter. To some extent, machine translation cannot flex with a number of adjustments, such as:
- Added/omitted words in the target;
- Wrong word order for compound nouns;
- Wrong gender or number;
- Wrong punctuation;
- Wrong capitalization;
- Words in source language left in the target; and
- Wrong/missing articles.
It is a good news that the improvement of neuronal machine translation has been leading to the betterment of post-editing quality in terms of consistency and efficiency through the utilization of technologies. With the fact that machine translation’s performance is never as flexible as human translation, the improvement of post-editing machine is now more massively pursued in the translation industry. The appearance of PEMT actually marks the rise of computer-based language services that quickly can handle large volume of texts from numerous language pairs. Combined with the linguists’ sensitive and capable skills, the existence of PEMT is expected to grow and stabilize the industry in the current global era, where language is considered as an urgent tool for a human-to-human interaction.
The quality of post-editing levels may vary based on either client or end-user. Thus, the utilization of PEMT will consider two types of post-editing features, including light post-editing and full post-editing. These features generally run through a productivity tests by machine translation users prior to the execution of the project in order to determine the effectiveness of the machine in managing certain language pairs.
The following diagram differentiates both types:
Light Post-Editing (LPE)
The light post-editing (LPE) modifies the raw outputs of machine translation to ensure that the quality can legibly and accurately convey the meaning of the source texts. With this type of product, post-editors are demanded to investigate any errors that might hinder the document’s purpose, or else the final results will be so much embarrassing results.
The full post-editing (FPE) can assist post-editors to thoroughly review and modify of the quality of translation products. This type of post-editing emphasizes the bare essentials of accuracy and legibility by considering a number of factors, including but not limited to stylistic and tonal consistency within the document, the absence of all grammatical errors, and appropriate cultural adjustments for the target language.
The source texts that have been scrutinized through the FPE process will be more convincing to the readers in the target language. Compared to LPE, FPE is more expensive, yet you can rest assured of its highly qualified outputs.
Post-editing, in fact, does not stand alone as the ultimate end of translation process. To produce a decent structure in the target language, translators (and editors) are firstly required to deal with pre-editing procedures, especially regarding with writing styles. The pre-editing steps should pay attention to the clarity of sentences and avoid any unambiguous segments to ensure smooth processing by MT. It is also substantial to ensure grammatically correct source to prevent from illogically production of text composition.
The utilization of active voices will also support the MT to work properly in cutting the vagueness in MT outputs. Other things to be checked include the probability of misspellings, complex sentences (which might be wordy), as well as clichés and colloquial phrases.
The Growth of Post-Editing Machine Translation
With a high quality of features, well-trained machine translation can approach the quality of human translations with a requirement of only a light manual touch for the refinement. Nevertheless, the quality of machine outputs should undergo the investigation with post-editors to determine the productivity goals and post-editing levels. It is important to note that for this type of post-editing, only major mistakes that impact the ability to understand the text are corrected, such as mistranslations, omissions, and additions. Meanwhile, the spelling, grammar, or punctuation mistakes may not always be corrected, in which can be a real challenge for any post-editors.
Post-editing might be a challenging responsibility, since it is the end tail of translation process that defines how refined and acceptable a certain translation product is. A certain degree of professionalism and operational experience should be considered as the priority to handover your articles to certain translation agencies. With a high credibility of its team, Linguwiz offers you friendly post-editing services that guarantee your document output quality. We work through efficiency at competitive service standard to provide you with a hassle-free solution. Kindly click the contact us button for more information.