The Internet and AI era have given the age-old art of translation entirely new meanings. Just as with the AI boom started by AlphaGo just a few years ago, after more than 10 years of rapid development, machine translation has also drastically changed the translation industry. While in the past, translation was considered an AI-incompatible cultural discipline that relied entirely on personal knowledge and experience, the industry is now focusing more on how to better integrate machine translation.
In terms of machine collaboration, while MTPE (machine translation post-editing) is already a well-known model in the industry, many don’t know that there is another “PE”, namely pre-editing. Despite the similarities of their names, pre-editing is significantly technically different from post-editing and has to potential to lead to completely different results.
What is Pre-Editing?
Though the purpose of both post-editing and pre-editing is to improve the quality of machine translation, post-editing focuses on improving the quality of the target text after it has been translated, while pre-editing focuses on polishing the source text before translation to improve the quality and accuracy of the machine translation. As an upstream link to the post-editing process, pre-editing can help refine the source text to significantly reduce the workload of post-editing, and in some cases, even render the process unnecessary to a certain extent.
What Does Pre-Editing Entail?
So just what are the criteria for the pre-editing process? As the ultimate goal for this process is to improve the quality of the resulting translation, the criterion for judging its success should not be the quality of the edited source text, but rather how easy it is to machine translate. In other words, those response for pre-editing first need to understand the complexities of the machine translation process and use this knowledge as a guide to optimizing and reforming the source text in a manner more conducive to machine translation. This includes:
- Correction of typos, punctuation, grammar, and other basic errors in the source text.
- Optimizing the use of terminology, correcting improper and unstandardized terminology, and ensuring consistency in the source text.
- Rewriting and simplifying unnecessarily complex sentence and grammatical structures in the source text, and breaking down long sentences into shorter ones.
- Eliminating any possible ambiguities in the source text, while replacing idioms and slang with simple and straightforward expressions.
The pre-edited source text should ensure a more accurate and clear target text after machine translation, one that requires only a small amount of coordinated post-editing to be viable for delivery.
The Unique Value of Pre-Editing
Both pre-editing and post-editing provide key value to the machine translation business model. Let’s take a deeper look at the unique value that pre-editing offers.
Pre-Editing Costs are Lower: As post-editing requires competency in the target language, the personnel and time required rises with the amount of target languages in the MTPE project. Pre-editing, on the other hand, only requires a single pass from editors competent in the source language.
Pre-Editing is More Efficient: Consider a scenario in which there is an obvious error present in the source text. Once machine translation is carried out, this error must then be identified and corrected by the post-editor of each individual target language. With pre-editing, this error needs only to be identified and corrected once, thereby greatly increasing the efficiency of the overall project.
The concept of pre-editing has in fact existed since the beginning of the language services industry, listed under the umbrella of “pre-processing/engineering” by most professional LSPs. The scope of pre-processing depends on the specific format, but often includes language-related elements to improve the efficiency and quality of the translation process.
Both Pre-Editing and Post-Editing are Important
Though translation has been a controversial topic since its introduction to the industry, MTPE, serving as a key component of human-machine collaboration, has greatly improved its practicality and enabled its deployment in a wider range of scenarios. With the popularization of post-editing applications and greater emphasis placed on upstream businesses, pre-editing and post-editing have been integrated into a cohesive whole to achieve the best results. The pre-processing or pre-editing of the source text in accordance with specific standards can achieve obvious results in terms of quality, efficiency, and cost reduction regardless of whether you’re planning to use machine or human translation. In other words, both pre-translation preparation and post-translation optimization are equally important for any translation project.