Check out all the on-demand sessions from the Intelligent Security Summit here.
Seeking to target enterprise customers with AI language translation, Cologne, Germany-based DeepL announced a new funding raise that public reports estimate at well over $100 million. Language translation is an increasingly critical function for enterprises working across geographies and different demographics.
Basic language translation capabilities have been available on for decades — for example, services such as Google Translate. But the challenge has been enabling more advanced translation for business use cases that capture not just the literal meaning but the right tone and context. This is an area where AI powered language translation is beginning to make an impact.
DeepL launched in 2017 and has steadily advanced its technology through deep neural networks. The new funding raises the company’s valuation to more than $1 billion. The company did not publicly release the total raised.
“We are not disclosing this number, we can just say that it is of significant size,” DeepL CEO, Jaroslaw “Jarek” Kutylowski told VentureBeat.
Intelligent Security Summit On-Demand
Learn the critical role of AI & ML in cybersecurity and industry specific case studies. Watch on-demand sessions today.
While coy on what the actual funding amount, Kutylowski has very clear objectives on what the money will be used for. As the company is growing, he noted, it will be spending more on fundamental AI research, venturing into new product areas and also expanding its portfolio towards enterprise customers.
AI-powered translation is a growing trend
The earliest days of language translation were driven by basic pattern matching techniques.
For example, a user types “hello” into a database that matches its equivalent in another language; for example, French (“bonjour”). The basic semantic constructs of pattern matching, however, don’t scale for larger scale translations, where context and tone matters.
DeepL is one of many vendors that have been applying advanced AI techniques to better translate human language. Google has been advancing its Google Translate service in recent years with a series of different approaches, including the use of a recurrent neural network (RNN).
Microsoft has been actively updating its Azure Translator service with AI models that the company claims improves overall quality. Meta (fomerly known as Facebook) isn’t being left out of the party either, announcing its AI powered universal speech translator (UST) project in October 2022.
Taking a deep neural network approach to language translation
DeepL has developed a language translation engine that relies on neural networks (NN) to infer accurate translations.
According to the company, it uses a novel NN design to understand the nuanced interpretations of phrases and sentences and is able to convey them in a target language.
“We do not disclose the details of our translation technology, but can say that as a company we’ve always been pushing the boundaries of how neural networks are designed to maximize translation quality,” Kutylowski explained.
The original core vision of the company was to break down language barriers — and Kutylowski emphasized that the company continues to focus on this area.
“In the beginning we understood this vision as being very strongly tied to translation specifically,” he said. “As we develop further as a company we see us using the underlying technology to help humans to communicate also in other ways — with new products that facilitate communication.”
The continuing enterprise challenges of translation
There are numerous challenges that enterprises face when dealing with translation that DeepL is looking to help solve.
Kutylowski commented that the world is becoming more interconnected every year, which serves to increase the importance of language translation and communication. In his view, with that growing demand, approaches to localization might be too slow and are just not able to scale.
“Tools like DeepL allow the end user, whether in the marketing, legal or any other team, to communicate across borders and publish material directly, without the need for reaching out to specialized teams or hiring,” said Kutylowski. “That creates totally new opportunities.”
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.
Author: Sean Michael Kerner