I am translating a passage and using Neural Machine Translation. NMT seems to override Termbase, and it disregards the fact that some terms should be translated in a certain way as I instructed Termbase to do. Isn't it possible for Termbase to kick in automatically while using NMT? Otherwise I don't see the use of it, if I have to correct NMT's mistakes manually.
Hi Lydia, the article you attached addresses a different issue, and it involves the use of "SDL MT Cloud", which is not even present in the list of machine translation tools in my Trados Studio 2021 software (the only one that seems to be working is "Language Weaver"). A Trados support representative wrote to me this morning saying that "NMT overrides Termbase", which means that the latter is pretty useless, since most translators nowadays work with machine translation (combined with TM and glossaries, i.e. "Termbases"). I need all of the three features (NMT, TM, Termbase) to work together. NMT and TM seem to work together, but Termbase doesn't seem to. It only suggests a translation when NMT is switched off (which means I'm translating manually, like in the pre-machine translation era).
Hi Lydia, the article you attached addresses a different issue, and it involves the use of "SDL MT Cloud", which is not even present in the list of machine translation tools in my Trados Studio 2021 software (the only one that seems to be working is "Language Weaver"). A Trados support representative wrote to me this morning saying that "NMT overrides Termbase", which means that the latter is pretty useless, since most translators nowadays work with machine translation (combined with TM and glossaries, i.e. "Termbases"). I need all of the three features (NMT, TM, Termbase) to work together. NMT and TM seem to work together, but Termbase doesn't seem to. It only suggests a translation when NMT is switched off (which means I'm translating manually, like in the pre-machine translation era).
Hi Alex Dall'Asta, no the link I sent is relevant because you have a termbase which you want to use alongside NMT.
For this, the termbase you have need to be migrated into a dictionary. A dictionary alongside NMT will give you an output that is customised based on what you have in your dictionary (termbase).This will reduce the amount of post editing you are doing. However looking into your account for NMT, you currently dont have dictionary support.
So this is what you need to do:
1. Log into: https://portal.languageweaver.com/settings/account
2. Upgrade your current plan to at least starter - so that you can have dictionary support
3. Export your termbase so you can import into your dictionary
Help on this can be seen here: https://docs.rws.com/791593/720646/language-weaver/dictionaries
Further support can be obtained from the team direct by using the contact us option
Lyds
Lydia Simplicio | RWS Group
_______
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
I came upon this discussion when trying to find exactly the same information as Alex. Unfortunately, the link to the article you refer to no longer works. Can you provide an update?
Mats Linder
Please try a different browser as https://portal.languageweaver.com/settings/account does work.
Alternative would be if you are in the USA,
However you sign in, you need an account.
Lyds
Lydia Simplicio | RWS Group
_______
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
You misunderstand -- I was not asking about Step 1 in the instructions you gave Alex. I was asking about the article that you directed Alex to (how-best-to-handle-differing-client-terminology-when-working-with-neural-machine-translation), and that link leads to a Not Found result.
Translate
