ULG’s Language Solutions Blog

Takeaways & Trends: What We Learned at LTIS 2019

Posted by Deborah Contreras

 Success in the language services industry isn’t possible without pushing the boundaries of technology.  

The Language Technology Industry Summit (LTIS) is a European conference that convenes on multilingual ambient intelligence to explore new developments in areas such as speech interaction, deep meaning processing and communication and cognition.  

Our attendance at this year’s LTIS was laden with thought-provoking insights that span how the language business model is being upended and which trends are promising to change the industry as we know it. 

Key Takeaway: It’s time to revise the traditional translation model. 

As today’s Neural Machine Translation (NMT) output continues to become more sophisticated, the roles of subject matter experts, translators and linguists are changing. 

That doesn’t mean the traditional translation business model is going away completely, but it suggests adaptation and reworking are needed to stay relevant. With technology innovations like NMT, Language Graphs, automated domain classification and more – sticking to the old way of operating won’t cut it. 

An example of how language partners, like ULG, are updating the translation model is in the workflow of linguists and NMT. Now, linguists work heavily on the front-end to prepare and enrich NMT data, rather than post-editing the NMT output. They also are an integral part of building the terminologies, taxonomies and Language Graph systems to ensure the quality of materials translated with NMT. 


Trend Outlook: The tech of tomorrow is here, but NMT continues to be a mainstay. 

At LTIS, we explored how technologies that were once musings of the future are now becoming reality. Technology considered nearly impossible like Real-Time Speech Translation is now at the doorstep of the industry. Another trend at the forefront of today’s language technology landscape is data-mining in machine-readable lexicon information for Taxonomies and Language Graphs, with data that is linkable to domain-specific corpora and other resources. 

While these new breakthroughs develop, NMT is a trend that is here to stay and a commodity that every language service provider (LSP) needs to have in its roster in order to compete. The challenge for many LSPs is the complexity of NMT makes it difficult to master, which is crucial to its successful implementation. If an LSP does not have a strong NMT system, the result of their translations for clients will suffer. Building robust technology that is flexible, fast and accurate is the key to optimal NMT, and an area we know a thing or two about at ULG. 


Interested in learning more about technology that is changing the language industry? Download our eBook. 

[et_bloom_inline optin_id="optin_3"] Google+ Facebook LinkedIn Twitter Instagram