memoQ amp SYSTRAN Partner Webinar Learn the Ins and Outs of Neural Fuzzy Adaptation

>> YOUR LINK HERE: ___ http://youtube.com/watch?v=KaLQHRFXIDI

Featuring more than 55 languages (and 150 language combinations), SYSTRAN is one of the most renowned neural machine translation providers. In this webinar, we will walk you through how to use the SYSTRAN-memoQ integration, focusing on the new Neural Fuzzy Adaptation functionality, introduced in memoQ 10.0. • TIMESTAMPS: • 00:00 - Introduction • 01:58 - About memoQ • 02:52 - About SYSTRAN • 06:45 - NFA in a nutshell • 19:00 - NFA in memoQ • 22:55 - Q A • 23:18 - Is the system able to SWITCH between desired styles? i.e. in medical translation between patient friendly language and documents aimed at healthcare professionals • 24:01 - Does NFA work when you do a pre-translation with MT, or does it only work during live translation? • 25:10 - Does the NFA get triggered during the translation request? And is so does it add processing time to the overall request result time? • 26:17 - If I understood correctly the NFA is uses the TM's that were not available before if we are talking about translating content with a custom model that was trained on clients content before? Is this statement true? • 27:27 - What is the minimum recommended fuzzy match to be useful to NFA? • 28:50 - How does Systran protect security of confidential document content? • 29:23 - Are the data of the post-edited sentences stored by Systran in their engine? • 30:30 - Is NFA capable of machine learning and improve the TM as long as it is used? • 32:57 - Is NFA offered as part of the basic Systran model, or is it offered as an additional service package? • If there is more than on TM with high match for the processed segment, 33:35 - how the system chose which TM should be used for NFA • 34:48 - Will it be possible in the future that SYSTRAN learns directly from the corrections in memoq, instead of having to retrain the engines regularly? • 35:30 - Is there a limit on how many TMs the NFA can work with in one project? • 36:05 - How does the formatting adaptation work? Do you have to define rules, if yes, where? • 37:20 - Is NFA only available in the newest version of memoQ? Or is it backwards compatibile? • 37:47 - Apart from the clear application to fuzzies, is the general idea of NFA to teach the generic MT in order to improve the quality of MT applied on no match words? • 38:50 - And what if there is 3 TM with the same match? • 39:32 - What about pricing? Do I need a monthly account, or is there also a pay-as-you-go account? • 40:36 - Are there considerations on memoQ's side to also send context to improve the NMT (for example the last x segments and their translations)? • 41:24 - Is the ability to enable or disable NFA via the API? How about the memoQ API

#############################

New on site