PEX is a technology used to post-edit the translations generated by a KantanMT engine. PEX can reduce the amount of manual post-editing required for a particular translation, reducing project turn-around times and costs.
Using PEX means you can avoid the costly and time-consuming post-editing step, especially in highly repetitive, highly formatted documents.
Post-editing involves the correction of machine translation output to ensure it meets a level of quality. There are two broad types of post-editing :-
Light post-editing (sometimes referred to as Rapid post-editing) aims to make the translation moreunderstandable while ignoring all stylistic issues. Common light post-editing tasks include replacing unknown words, removing redundant words, fixing machine-induced meaning distortion, and partially or completely rewriting sentences.
Full post-editing aims to make the translation stylistically appropriate. This degree of post-editing is most similar to standard editing and involves all of the practices that a translator/editor follows including checking terminology, making grammatical and syntactic adjustments, adjusting for target language fluency etc.
PEX is used in both of these post-editing scenarios to ensure that the need for post-editing is kept to a minimum during a translation project.
In the scenario below, product ID numbers have been incorrectly cased throughout the translation. This has created several hundred instances of incorrectly formatted product IDs that will need to be corrected during the post-edit phase.
Source Text: Modèle SP17 pour ordinateur 17"
Target Text: Modelo sp17 para computador 17”
Traditionally, a post-editor would be tasked with reviewing these segments and modifying the target text, a process that can be both time-consuming and expensive.
However, a single PEX construct can be used to automate the correction or post-editing of all product IDs.
In this example, the letters ‘sp’ followed by a minimum of two to a maximum of four digits are matched and replaced with ‘SP’ followed by the same matching digits. This single PEX construct will now correctly format every product ID, ensuring that a translator can ignore these post-edits entirely.