Artificial intelligence is just getting started in the translation business, which means a variety of different approaches are currently being taken to glean maximum profit from new megadata-based tech. Neural machine translation (NMT) has been a resounding success and the quality output it produces is no doubt impressive. That said, these texts require postediting by a real live translator, who is often better equipped to resolve certain issues compared to their computer counterparts, especially when said glitches are contextual.
At Terminotix, we believe neural machine translation poses no threat to human translation—it’s actually an asset. An NMT system can be a key tool in a translator’s arsenal. Read on to learn just how each of our products interact matchlessly with your preferred neural machine translation solution.
NMT system output quality is only as good as the data it’s trained on. That’s great news for translators, who have large quantities of high-quality data at their fingertips in the form of translation memories. AlignFactory puts several alignment filters and formats at your disposal (XML, HTML and TMX), meaning you get compatibility across all major NMT systems. That also means when you use AlignFactory to automate text alignment, you’re also seamlessly entering a document corpus into your neural machine translation solution and training it at the same time.
Used in tandem with more generic text corpora, a rich termbase is a vital asset when it comes to training any NMT system. SynchroTerm allows you to efficiently feed a termbase, which can in turn be imported into your NMT system and enable it to produce specialty-specific translations. What’s more, if you need to meet varying needs all at once, you have the option of creating different profiles linked to particular areas of expertise.
Our translation environment lets you integrate SYSTRAN and Portage NMT systems into a single pre-translation tool: LogiTrans. Plus, you can share LogiTrans with your associates through the Microsoft Word LogiTerm toolbar. To boot, online NMT resources like SYSTRAN, Google Translate, DeepL and Bing can be integrated into the interface as external search engines, allowing you to use them without ever leaving LogiTerm.
With Croatian, Estonian, Gaelic and Maltese up and running, SynchroTerm now serves all the official languages of the European Union. Go to the Terminotix site for the full list of all thirty languages offered.
Several new features have been added to make using SynchroTerm even more seamless.
Now you can use SynchroTerm to only extract certain pre-set terms from documents. All you have to do is import a list of terms using the Restrictive Terminology function before you kick off term extraction.
Two separate options now allow you to choose between simply removing terms from the list of extracted terms and marking terms as “noise”. Marking terms as “noise” is a snap via a button in the Extraction window, allowing SynchroTerm to progressively gain efficiency as you use it. What’s more, the list of terms to exclude is fully modifiable by way of a utility that lets you import, export, delete and add terms as well.
A new feature allows you to refine searches using translation equivalents derived from bitexts. When the Translation spotting option is turned on, LogiTerm generates an extra column where each equivalent is listed with its number of occurrences (located on the right-hand side of the results page). You can check results by equivalent or remove certain equivalents from these same results.
The Académie française recently launched a search engine to facilitate access to its dictionary. Its integration into LogiTerm is under development.
In recent months, we’ve made a bunch of improvements to BridgeTerm Web:
McGill University’s School of Continuing Studies is offering up an all-new specialized course in legal translation that covers the field of securities law. Open to translation students and practising professionals, this course aims to prepare translators to fill an important gap in the market. Dates: February 21-April 11, 2019.