Avilinga ParTex efficiently aligns translation memories for more effective computer-assisted translation (CAT) and machine translation (MT) tools.
If you possess previously translated content but no translation memory (for instance, if you’ve been translating without using specialised translation tools), we can create one for you using ParTex.
What is a translation memory?
A translation memory is a record of all previous translations. They are fundamental to machine translation, but can also help linguists work more efficiently by identifying parts of the source language that have already been translated into the target language.
Building translation memories begins with matching sentences in the source language with their equivalents in the translations. This is called translation memory alignment.
Computers can only form an approximate alignment based on sentence order (i.e. matching sentence one in the source with sentence one in the translation, then sentence two with sentence two, and so on). But quality translations often require affecting sentence order, and even splitting or merging sentences. That means the computer’s alignment output must always be reviewed and corrected by a linguist.
Until the creation of Avilinga ParTex, this review process was a laborious one. ParTex is a machine-translation based translation memory alignment tool that enables easy review and correction of computer-generated alignments.