Hello Yumiko SHINDO,

The uplift technology what it does is that leverages segment fragments from TMs to increase productivity and speed.

Uplift is available through 2 new functions

• Fragment recall – with this Studio identifies matching fragments from a TMs existing TUs, so fragments in new segments are matched against whole TUs and fragments in new segments are matched against corresponding fragments inside Tue .
• Fuzzy match repair – improves the TM fuzzy matches by looking for replacements for mismatched source content in other resources .

Uplift fragment matches is not enables by default and once the TM surpassed 1000 units  you can use the uplift functionality (the minimum for upLIFT to works is 1000 units )

Also you can now use uplift features with TMs that do not have the same language pair as the project you are adding them to as both upLIFt fragment recall and Uplift match repair are avaulable with TMs you add to your project with AnyTM

Fragment alignment support is now available for Asian languages and enables users with Chinese or Japanese source or target languages to make use of upLIFT Fragment Recall and upLIFT Match Repair for the first time.

For upLIFT Fragment Recall and Match Repair to work for Chinese and Japanese, we changed the way Trados Studio identifies words in these languages. If until now, text segmentation for Chinese and Japanese was done on a character to character basis, this has changed with Trados Studio 2017 SR1 so that words are now the basic unit for segmentation and not characters. This helps Trados Studio extract meaningful fragments for the upLIFT features to work with.

Implications in word counting and fuzzy matching

To enable the innovative upLIFT functionality, we moved to a different way of attempting to identify words in Japanese and Chinese source languages. This implies that both fuzzy matching (including metrics in analysis reports) and word counting have changed. As a result, word-based Analysis Statistics and Confirmation Statistics may provide different results than before and you might also get different leverage from your TMs. You will still be able to find character count in the reports.

These changes are meant to give a more accurate view of the work that translators of Chinese or Japanese need to perform.

Hope this can help you.

Best regards,

Ana-Maria Matefi

Hello Yumiko SHINDO,

The uplift technology what it does is that leverages segment fragments from TMs to increase productivity and speed.

Uplift is available through 2 new functions

• Fragment recall – with this Studio identifies matching fragments from a TMs existing TUs, so fragments in new segments are matched against whole TUs and fragments in new segments are matched against corresponding fragments inside Tue .
• Fuzzy match repair – improves the TM fuzzy matches by looking for replacements for mismatched source content in other resources .

Uplift fragment matches is not enables by default and once the TM surpassed 1000 units  you can use the uplift functionality (the minimum for upLIFT to works is 1000 units )

Also you can now use uplift features with TMs that do not have the same language pair as the project you are adding them to as both upLIFt fragment recall and Uplift match repair are avaulable with TMs you add to your project with AnyTM

Fragment alignment support is now available for Asian languages and enables users with Chinese or Japanese source or target languages to make use of upLIFT Fragment Recall and upLIFT Match Repair for the first time.

For upLIFT Fragment Recall and Match Repair to work for Chinese and Japanese, we changed the way Trados Studio identifies words in these languages. If until now, text segmentation for Chinese and Japanese was done on a character to character basis, this has changed with Trados Studio 2017 SR1 so that words are now the basic unit for segmentation and not characters. This helps Trados Studio extract meaningful fragments for the upLIFT features to work with.

Implications in word counting and fuzzy matching

To enable the innovative upLIFT functionality, we moved to a different way of attempting to identify words in Japanese and Chinese source languages. This implies that both fuzzy matching (including metrics in analysis reports) and word counting have changed. As a result, word-based Analysis Statistics and Confirmation Statistics may provide different results than before and you might also get different leverage from your TMs. You will still be able to find character count in the reports.

These changes are meant to give a more accurate view of the work that translators of Chinese or Japanese need to perform.

Hope this can help you.

Best regards,

Ana-Maria Matefi

Thank you for your answer.

Today, I participated in the SDL roadshow in Tokyo and I got a occasion to ask directly this question  to a SDL staff.

The response is No.

The languages pairs French - Japanese / Japanese - French are not covered by upLiFT.

This SDL staff told me the possibility to translate texts at first Japanese to English and then English to French using upLIFT, but it’s not so appreciated because I saw sometimes these two steps process of translation  would change some meanings of original texts....

Anyway, now I know at least that my languages pairs are not covered by upLIFT so just I try to find alternatives.

It’ll be highly appreciated if you could give me some suggestions.

Thank you for your answer.

Best regards,

Yumiko SHINDO 

Hi Yumiko SHINDO

Thank you for attending SDL RoadShow.

I think you asked about LauguageCloud but not upLIFT then.

If you meant upLIFT, it supports French and Japanese.

LanguageCloud is Machine Translation by SDL. This does not support language pair French - Japanese.

upLIFT is a new technology of Translation Memory. This can handle language pair French - Japanese.

Hi there,

Yes, I can confirm that you asked me as well about JA<>FR language support for BeGlobal MT, to which I said no, it is not currently available.