PEMT (post-editing machine translation) is a feature that we implemented a few months ago in some language pairs as a way to leverage technology to make the translation process more efficient while increasing consistency and quality.

In jobs with PEMT enabled, the target section is pre-filled with a machine suggestion, which we expect you to use as a reference to make appropriate edits when necessary. As always, your resulting translation must be accurate and read naturally in the target language. Here are some tips for producing natural translations.

 

Here’s how to work on collections with PEMT enabled:

You can identify jobs or segments with PEMT by a ‘MT’ icon on the top left corner of the target box, which will disappear once you click on that job.

pemt-1.png

On the right panel (open by default, or by clicking the light bulb icon), you will see MT suggestions, as well as glossary terms and TM matches when present, and clicking on any of them will copy the contents into the target box.

pemt-with-tm.png

Once you accept the collection, each job will have a green “Confirm” button that you will need to click before you can submit, ensuring that you have reviewed the MT suggestions and edited them when necessary. You can use keyboard shortcut Ctrl + Shift + Enter to speed things up, and the cursor will move to the next job.

pemt-confirmed.png

Keep in mind that any {1}deflated HTML tags{/1} or content [[[inside of triple brackets]]] may still be picked up by the machine translation, causing errors (highlighted in red) that won't allow you to confirm or submit.

pemt-triple-brackets.png

You will need to fix those errors in HTML tags or triple brackets before you can confirm these jobs.

pemt-triple-brackets-fixed.png

To keep track of your progress, you can use the filter button to easily find "Confirmed" and "Unconfirmed" jobs.

pemt-filter.png

If you need to fix a job you have already confirmed (but before submitting everything), you can still edit it by selecting “Unconfirm” under the “Confirmed” button drop-down.

pemt-unconfirm.png

Once all jobs are confirmed, the submit button will turn dark green and clickable, and you'll be able to submit the whole collection at once.

pemt-all-confirmed.png

When a job is returned for corrections, you can find it under the "Revising" filter (as usual and as explained here), but you will need to click "Unconfirm" before the job can be edited. 

pemt-revising.png

 

 

FAQ:

How does PEMT help the customer?

It can improve consistency, reduce costs and improve speed of their translations.

How does PEMT affect my rewards?

The rewards for jobs with PEMT enabled are automatically adjusted, and are paid at 75% of the original rate. Example: $1.00 > $0.75

What’s the difference between PEMT and TM?

TM matches are translations that have been done previously by another translator, and then reused to aid you on your translation work.

On the other hand, PEMT suggestions are translations done by a machine, with no human input, and are meant as a reference.

What if a job has both PEMT and TM?

In that case you will be able to choose between PEMT suggestions and TM matches. The job may have one of them prefilled, following this level of priority:

  • TM matches will be prefilled if it has 100% or 101% matches (75 to 99% TM matches will not be prefilled).
  • If TM matches are less than 100%, PEMT suggestions will be prefilled instead.
  • When clicking on the “matches” (light bulb) button, you will see Glossary terms at the top, then 101% TM matches followed by 100% TM matches, then PEMT matches, and lastly 75 to 99% TM matches. 

Where do PEMT suggestions come from?

They are generated by Google Translate.

Who is responsible for bad PEMT translations?

Translators are responsible for ensuring that they only submit accurate translations, so you should edit MT suggestions as necessary. 

What if the MT suggestion is not usable?

Make sure to flag jobs where the initial suggestion is not useful and would not help you save time on your translation. You can also contact Support for further assistance.

What if the MT suggestion is already good as it is? Will I get in trouble for submitting MT?

As long as you are submitting good quality translations, there is no problem in submitting MT suggestions directly if you feel they don't need any edits. Remember all translations you submit must be accurate and read naturally, and achieve a minimum GoCheck score of 7.

Are there any keyboard shortcuts to help make using MT more efficient?

Yes, there are some useful shortcuts, and we will be adding more in the future so please let us know what would be helpful to you.

View matches and suggestions: ctrl + shift + m

Insert the top match/suggestion: alt + shift + m

Confirm edits: ctrl + shift + enter

Will Gengo consider adjusting PEMT rewards based on our feedback?

By introducing PEMT, many new customers have started working with Gengo who otherwise wouldn’t have chosen us due to their specific need for this feature. These customers have high volumes—and a consistent amount—of work that they are giving our translators across many language pairs. Because of PEMT, we have even opened up new language pairs that didn’t previously exist. Thus, giving new opportunities and work to Gengo translators.

However, we do always aim to pay a fair rate to our translators for all work that is required, whether translating or reviewing a suggestion. As the volume of jobs using PEMT grows, we’ll continue to try and maintain a fair balance and will be constantly reviewing the situation. 

As demand for PEMT continues to grow, not only at Gengo but within the translation industry as a whole, we’re aiming to be able to bring on more new clients by extending the PEMT feature to a greater number of jobs ordered in your language pair. We expect that extending the PEMT feature to more collections will continue to help you make better use of your time, enabling you to access and pick up more work.

As always, please don’t hesitate to reach out to Support with any questions or quality concerns. We are working on creating a feature that will make the process of flagging poor suggestions smoother. However, for the time being, we’d appreciate your help in letting us know if you experience any MT, PEMT or Edit-related issues.

Have more questions? Submit a request

0 Comments

Article is closed for comments.
Powered by Zendesk