Despite the abundance of tools in the market, translation productivity (CAT) remains an unsolved problem. There are still gains to be made and the market remains fairly fragmented. A recent report from a number of EU institutions identified gaps in existing productivity tools in terms of user friendliness and interoperability and stated that “the future smarter CAT environment is yet to be developed.”
Indeed, the era of continuous localization and the advent of artificial intelligence have brought about fresh challenges and opportunities in the field of translation productivity.
On the sidelines of a recent customer event in London organized by SDL, developer of productivity tool Trados, Slator spoke to SDL CEO Adolfo Hernandez. The event featured a preview of their new translation management system (TMS), SDL Language Cloud, which is set to go live in September 2019. A number of Language Cloud modules include translation memory (TM), terminology management, automated workflow, neural machine translation (NMT), and a content analyzer.
Hernandez pointed out how “the dial of adoption is moving” as NMT usage in the language industry becomes widespread across LSPs and end-users alike. In fact, according to the aforementioned EU report, NMT has become “an integral part of a linguist’s toolbox,” and “most EU translation services […] provide NMT output to their linguists.”
Hernandez told Slator that “in general, MT is very pervasive in terms of how our linguists work.”
Beyond machine translation, there is potential for broader AI to be applied to other facets of language production. Speaking at SlatorCon San Francisco in September 2018, Hernandez shared his views on NLP, saying, “There’s a lot more to natural language processing than translation.”
SDL Language Cloud’s content analyzer is a foray into this wider application of AI. The content analyzer uses AI to identify the subject matter of the source content through tagging. This has the potential down the line to help automate project allocation by, say, matching a subject matter to suitable freelance translation resources. “SDL Language Cloud is the first solution to use ML to understand content before it is handled for translation,” Hernandez said.
Continuous Localization
Slator also spoke to Andrew Thomas at the same event. Thomas is Senior Director of Marketing at SDL. He said what makes SDL Language Cloud different is that “it’s being designed specifically for continuous localization.” Continuous localization requires on-demand localization at scale and, Thomas said, “that’s a scale that nobody has really tackled yet because the infrastructure wasn’t there.”
“In general, MT is very pervasive in terms of how our linguists work”
— Adolfo Hernandez, CEO, SDL
As the name suggests, SDL Language Cloud is a cloud-based solution. It is designed for companies to manage their end-to-end localization process and combines machine intelligence with integration into SDL Trados Studio, among others. SDL Language Cloud is based on a microservices architecture, meaning that customers can choose to switch available modules on and off as required. Customers can also request human translation services from SDL through Language Cloud.
All, All, All
Thomas explained that a number of SDL’s other tools “were developed for different use cases and different niche types”; meaning they are primarily geared toward serving a number of specific customer needs. “Arguably, WorldServer was predominantly focused on the large enterprises with large complex issues. MultiTrans was really designed for the regulated industry with security in mind. GroupShare was designed for the LSPs because it was cost-effective and aligned with Trados,” he said.
But SDL Language Cloud is targeted at a broad customer base including enterprise clients, third-party LSPs, and translators. “The shorthand vision is ‘all all all’ — all content types, all translation methods, all users involved in the process. That is the goal,” Thomas said, adding, “Initially, we’re going to be focused on new customers. And we’re probably going to be focused squarely on localization departments within the enterprise.”
SDL is also “experimenting with pricing and packaging, in general, with this solution because we can turn things on or off very easily,” Thomas said. He confirmed that “this initial release is not radically different in pricing than some of our other TMS offerings.”
Meeting the diverse needs of enterprise customers, third-party LSPs, and translators is no small undertaking. Yet, as a translation technology provider cum large LSP, with some 1,400 internal linguists on the payroll, SDL already has a foot in multiple camps.
Slator asked CEO Hernandez for his view of the language technology landscape. According to Hernandez, “There are a lot of good people out there doing cloud TMSs. There are a lot of people who’ve got productivity tools and workbenches. There are a lot of people who’ve got neural solutions. There are some people who have project management and workflow management, but I haven’t seen anyone who’s got all of this integrated in the same place, as we have.”
Your information will not be shared with third parties. No Spam.