What’s Really Happening with MT? Highlights from MT Summit 2019
With 90% of the data on the internet created since 2016, global brands looking for international growth continue to integrate smarter language automation solutions. Technology-driven language solutions intelligently (and cost-effectively) handle the huge amounts of information and content published every day, without compromising quality.
Welocalize took part in the MT Summit in Dublin, contributing to several sessions looking at advancements and use cases of machine translation (MT) and neural MT. From those first dipping their toes into the world of MT to localization managers, translators, and professors, the focus at MT Summit 2019 was on how enterprises manage and process these high volumes of content and how to apply MT including NMT architectures to real commercial applications and scenarios.
“Welocalize processes hundreds of millions words per year using neural machine translation across different key verticals, including technology, legal, and travel. While NMT is delivering the highest quality in the history of NMT, Welocalize continues to push for more innovation.” Olga Beregovaya, VP of Language Solutions, Welocalize
Advancements in NMT + Post-Editing
Looking at the latest NMT architectures and customization approaches, the Welocalize tutorial, delivered by Welocalize MT dream team Lena Marg, Alex Yanishevsky, Natasha Latysheva, Anna Pizzolato, and Karen McLaughlin, focused on NMT and post-editing. This team works daily with global brands, developing advanced MT programs and marrying academic and research advancements with global business goals.
As more global brands require more speed and scale, the role of NMT and post-editing within enterprise MT programs becomes even more relevant. Advancements in customization, often using glossaries and Translation Memories, allows neural networks to deliver more accurate output which is important for certain vertical sectors, especially where content is regulated. Increased post-editing in the MT workflow is now common, with translators playing a key role and evolving their skills and the Welocalize team shared experience on how to set up a post-editing team for an enterprise MT program, discussing fair compensation models, and what makes a good post-editor.
Our colleagues delivered a successful tutorial as part of #MTSummit2019 where attendees learned about the latest trends in neural #MT, post-editing, MT performance evaluation, and more. The tutorial was well received, one attendee even claimed it was worth the visit to Dublin! pic.twitter.com/7QYNw5io2y
— Welocalize (@welocalize) August 21, 2019
Driving Commercial + Academic Collaboration
MT buzz in the media is well-known but for true success, MT must impact tangible ROI and help brands go global and reach international audiences. Bringing together academia, research and development initiatives, as well as commercial teams is putting MT in action, progressing translation productivity, quality, and output. The Welocalize sponsored Ph.D. track brought together many different people!
— celia rico (@celiaricoperez) August 22, 2019
Applying MT Across Vertical Sectors
Certain sectors may have not viewed MT as a viable solution – because of the high translation quality levels required. Areas like finance, legal, and pharma and biotech industries now view MT as a key part of their multilingual content strategy. As MT has advanced, many organizations in these sectors are starting to implement MT into the enterprise localization program. Maria Nunziatini presented a case study on the financial services industry and how NMT can play a key role in increasing translation productivity whilst maintaining the same top-level quality, using teams of specialist SME translators and post-editors.
— Andy Way (@tarfandy) August 22, 2019
The MT Summit 2019 was packed with incredible insights and forward-thinking visions – thank you to everyone who shared and participated!