This article explains how to insert translation memory matches in machine translations, known as Translation Memory Insertion.
What is Translation Memory Match Insertion?
Translation memory match insertion allows Account Owners and Project Managers to configure the percentage of translation memory (TM) leverage that can be automatically inserted by a Machine Translation (MT) Profile as its translation. The matched strings skip the translation step entirely, accelerating the translation process. Furthermore, the automatic insertion of TM strings saves translation costs associated with the MT Profile.
How it works
Setting a threshold of 80% means that all TM matches of 80% or higher will be inserted as translations in a workflow translation step which uses the MT Profile. Any strings with TM matches of 79.9% or lower are machine translated by the specified MT Profile.
Translation Memory
The translation memory used by the MT Profile in a Job depends on the project the content is translated in. The MT Profile leverages the translation memory assigned to the linguistic package in the project. This ensures the MT profile uses the correct translation memory within the project it is translating in.
Workflows
TM Match Insertion is currently only available on workflow steps that are configured with an MT Profile as the translation provider. It is available to use with any MT Profile, except copy-source-to-target workflows.
How to enable Translation Memory Match Insertion
- Go to Account Settings > Machine Translation Settings
- Click the Profiles tab
- On the MT Profile you want to enable TM Match Insertion on, click the "Disabled" link under the TM Match Insertion column
- Switch the toggle On to enable TM Match Insertion
- Configure the TM Match percentage that should be leveraged by the MT Profile
- TM matches that are the same or above the threshold are automatically inserted as the translation and skip the MT translation step
- The rest of the strings are translated by the MT Profile
It is highly recommended to enable TM Match Insertion only on MT Profiles which are used for workflows with a human post-editing step. This is to ensure that the inserted matches from your Translation Memory can be checked and edited by a human linguist. If Translation Memory matches are inserted without further post-editing, there may be a significant discrepancy between the existing translation from your TM and the new text to translate.
Recommended: Enabling Fuzzy Match Settings for post-translation steps
TM Match Insertion can help you achieve cost savings for Machine Translation workflows which contain a step for human Post-Edit, Edit or Review ("MTPE workflows").
Instead of a machine translation, the human editor will be able to revise the inserted Translation Memory match. Since this can significantly reduce the required editing effort, many Language Services Providers grant cost discounts for content with a fuzzy match.
To benefit from cost savings based on fuzzy matches, ensure to apply your Fuzzy Match Settings also to the post-translation step.
From Account Settings > Fuzzy Match Settings > Workflow Steps, select one of the following options:
- Apply your Fuzzy Match Profile settings to the Translation step and Post-Edit step:
This option will only take effect for MTPE workflows using the workflow step type "Post-Edit". - Apply your Fuzzy Match Profile settings to the Translation step and all post-translation steps:
This option will take effect for workflows using the workflow step types "Post-Edit", "Edit" or "Review", including workflows for human translation.
Important Considerations
TM Match Insertion does not insert TM matches if there is a placeholder mismatch between the source and the translation. In such cases, the system falls back to machine translation, and an issue is opened for the string. The failed TM match translation is recorded in the string history.
TM Match Insertion relies on segment-level matches. If a string has multiple segments, TM Match Insertion will insert any segment-level matches it finds. If a match is not found for a particular segment of the multi-segment string, the segment will be translated by MT.