Background Search for Terminology in STAR MT Translate

Background Search for Terminology in STAR MT Translate Giorgio Bernardinello STAR Group - Wiesholz 35 8262 Ramsen Switzerland giorgio.bernardinello@star- group.net Judith Klein STAR Group - Wiesholz 35 8262 Ramsen Switzerland judith.klein@star-group.net Abstract When interested in an internal web appli- cation for machine translation (MT), cor- porate customers always ask how reliable terminology will be in their translations. Coherent vocabulary is crucial in many aspects of corporate translations, such as documentation or marketing. The main goal every MT provider would like to achieve is to fully integrate the customer’s terminology into the model, so that the re- sult does not need to be edited, but this is still not always guaranteed. Besides, a web application like STAR MT Translate allows our customers to use – integrated within the same page – different generic MT providers which were not trained with customer-specific data. So, as a pragmatic approach, we decided to increase the level of integration between WebTerm1 and STAR MT Translate, adding to the latter more terminological information, with which the user can post-edit the transla- tion if needed. 1 STAR MT Translate STAR MT Translate is a highly customisable web application for machine translation (MT). It is not designed to be part of an automated translation process, nor to be a tool for expert translators, but rather to help any employee of a company understand texts and documents written in foreign languages. The UI can be designed to fit the corporate style of the client and it offers easy © 2022 The authors. This article is licensed under a Crea- tive Commons 3.0 licence, no derivative works, attribution, CC-BY-ND. 1 WebTerm is the STAR web-based terminology applica- tion. access to the STAR MT engines, specifically trained for each customer, as well as for the most well-known online MT providers. 2 TermAssist In the last few years, we have seen many companies starting to offer connections between terminology and MT, like the dictionary in GoogleTranslate or the glossary in DeepL. STAR started working almost five years ago on an integrated solution, where corporate terminology can be retrieved from the same webpage in which a text has been translated using MT.2 The purpose was to give the MT user the possibility of consulting the company’s dictionary without switching tabs in the browser, by simply highlighting one or more words. This function was named TermAssist and has since become one of the most requested functions of STAR MT Translate. The main limitation of this kind of approach is the lack of matches for inflected words. A very detailed dictionary could also contain inflections referring to the main term, but this is not realistic in practice, and it may get complicated for multi-word concepts. For example, the plural of the German word “Sitzplatz” (seat) is “Sitzplätze” (different vowel inside the word plus -e added at the end) while the corresponding Italian forms are “posto a sedere”, singular, and “posti a sedere”, plural (the last letter of the first word contains the inflection). 3 Background …

Content truncated to 500 words