5 months ago
June 23, 2021
RWS Revives Language Weaver, Now Powered by SDL and Iconic Translation Machines
Over the past five years, machine translation (MT) has rapidly matured to become an essential offering for global businesses. Many companies tout the simplicity of their MT solutions and, in many instances, generic technology meets clients’ needs. But for customers who need more, RWS is investing in research and expertise to go the extra mile with the newly revived Language Weaver brand.
Language Weaver’s return is a result of the collaboration between two heavy hitters in the MT world, Iconic Translation Machines and SDL, both of which RWS acquired in 2020. Although each company had its own unique way of approaching certain problems, there was enough overlap that the best approach was to merge the teams and their technologies.
Mihai Vlad, Managing Director of AI and Machine Translation at RWS, told Slator that the teams’ shared experiences have made for a smooth integration. “We’re surrounded by like-minded people who focus on similar challenges, like helping enterprise customers solve real business problems with machine translation,” he said.
The reintroduction of the Language Weaver brand signals a fresh start for RWS while honoring the teams’ heritage. “Far from starting from scratch, there’s a lot of history behind the two teams,” said John Tinsley, VP of AI Product and Marketing at RWS. “The name reflects that this is more than just MT at RWS.”
The name Language Weaver is a nod to MT pioneer Warren Weaver. His 1949 theory of computer translation as decoding and encoding languages was far ahead of its time, and aptly describes how MT functions today.
The company started off as a spin-out from University of Southern California in 2002, productizing research for business and becoming particularly popular with government and enterprise buyers. Iconic also has academic roots, emerging from Dublin City University in 2012 to eventually specialize in regulated industries such as IP, e-discovery, and life sciences.
Language Weaver will continue to tap into those strong university connections to support its research, a major priority for the brand. “Language Weaver is a destination for people who want to work on cutting-edge research. We want researchers graduating with PhDs in the field to think, ‘I want to be part of that team,’” Tinsley said, adding that research is not limited to product development, and goes beyond MT to touch on multilingual summarization and data analysis, among other areas.
“It could be that the MT requires other, similar NLP technology to get better,” Vlad noted. Combined teams, whose members include software developers, NLP researchers, and linguists, draw on their unique mixture of expertise to improve MT. This offers a unique position in the market, where “the group can either get the technology to enable productivity of linguists, or we use the linguistic prowess to make the technology better,” he explained.
Going the Last Mile
Language Weaver has already packed quite a punch in upgraded offerings for new and existing clients, and now handles more than 2,700 language combinations. And with plans to develop more services in response to buyers’ needs, the team is particularly driven by the goal of bringing MT to users wherever they might need it.
“A lot of times translation is more ad hoc than traditional localization,” Tinsley said. “You can have TMS and CAT tools that take care of industrialized translation as it’s going through the production line, but what about other cases where multilingual content is coming through email, in your CRM, or via a document review platform?”
In addition to Language Weaver’s integrations, which give clients access to MT within their preferred platforms, employees can send the machine translated content for linguist review with the click of a button, without leaving their workflow. Pre-agreed rates and other details simplify an otherwise complex process.
Language Weaver offers both off-the-shelf technology and bespoke solutions to help buyers go the last mile, which looks different for every client. For example, clients with highly sensitive data can opt to train models on-premise rather than in the cloud. The game plan might also include a hyper-customized engine, or post-editing and linguistic validation carried out by some of RWS’ thousands of linguists.
“We can get you most of the way there with our ready-to-go options, but when you need that last mile, that’s where we stand out and it’s one of the reasons I’m really excited to be part of this team,” Tinsley said.
Team members will share the history, technology, and linguistic expertise behind the brand when RWS’ six-part webinar series on Language Weaver kicks off on June 22, 2021.
The series will continue as familiar faces from RWS, SDL, and Iconic lead product demos and discuss enterprise solutions in several industry-specific webinars. Check out the Language Weaver blog for more details the Monday before each webinar.