“In over 50 years, the Machine Translation market has experienced several technological breakthroughs”

SYSTRAN is now 50 years old, but have you kept the soul of a “Start-up” and the ambition to be… a unicorn?

Vincent Godard – As a reminder, SYSTRAN was born with the beginnings of machine translation over 50 years ago. It initiated all major technological innovations in this field: the first free online translators (including Google), the first deployment of secure translation servers in large companies, the first machine-based translation AI with manufacturing technology and a unique language expertise.

If we consider our capacity for change, our agility and our company at the human level (115 employees), we can decent enough to say that SYSTRAN keeps its Startup soul. Today we are a specialist in automatic translation and we insist on our positioning as a pure player. Our clients are historically public organizations, government entities, but also large private sector accounts. Our only goal now is to provide the best quality machine translation for every user in complete safety.

In 2022, we have two main goals: to strengthen our activity with large companies and administrations in France thanks to the quality, security and confidentiality of data managed 100% from France and SYSTRAN assures them to develop our expertise in the cloud in that particular business. areas and thus expand our offer to 5,000 models already available; enough to also serve many uses in integration tools or in the validation of patents for example.

Today, we have many challenges and goals. Our latest turnover in 2021 reaches 16M €, with growth of 15% per year. So both confident and peaceful we are thinking about 2023.

However, more than being a unicorn, our first ambition is to find a public disappointed with current automatic translations, which are often too generic, who remain dissatisfied with current training machines and looking for an effective and efficient alternative. In this context, of course, we specifically target the professionals for whom we have developed adapted, customizable cloud services (for a special translation that allows us to go beyond what allows a generic translation now), with the ability to integrate across all systems. In a word: “tailor-made”.

For a very reasonable cost, all professionals have direct access to our cloud service dedicated to VSE-SMEs. SYSTRAN gives them the opportunity to translate a website into the language of their choice, and therefore grow their business across our borders. We now cover 55 languages ​​worldwide (from the most common to the most atypical), and we are present through our 5 subsidiaries located in Paris of course and which hosts our headquarters, San Diego in United States, Mexico City, Seoul in Korea or Japan in Tokyo.

I don’t know if these are the ambitions of a unicorn but one thing is certain, we have no intention of stopping on such a good path and fully intend to advance SYSTRAN around the world. The market is buoyant and growing, the crisis is behind us, we need to take home the point.

GAFAs launch free automated translation systems, sometimes in real time. How can Systran withstand this competition? By change?

Vincent Godard – GAFAs have certainly invested in the machine translation market. However, their free translations remain general, very approximate and unsatisfactory for millions of users, especially for all professionals or all those looking for a sharp, high-quality translation. They allow you to first understand the meaning of a sentence (and again, I’m referring you to the incident generated by the famous “God Bless You”, translated not so long ago as “God hurts you … by one of these digital players) but is in no way adapted to the languages ​​of professionals, or to the very specific translations associated with the trades they encounter.In other words, the generic (and therefore free) translation format corresponds to the general use of public.It may not be suitable for students, professionals, researchers, or anyone seeking translation at the expert level.

The particularity of SYSTRAN is to offer very detailed translation, which specializes in activity. We offer our customers a customized solution tailored to their businesses by using their internal terminology, their style and aligning ourselves with their quality standards. Herein lies the difference of our service. We also guarantee to all our customers the quality, security of their data and follow-up of a 100% French customer service based in Paris. And that’s not nothing when you know the universe of Gafam. Besides, I think I can just say we are.

Moreover, since you introduce me to the subject of change, I consider this to be key to our approach. We also regularly participate with our R&D team in Paris on many government-funded projects, all of which aim to optimize research projects around Artificial Intelligence or calculations.

We talk a lot about artificial intelligence but no one has really passed the “Turing test”. Do you expect technological breakthroughs, perhaps in quantum computing, that will finally allow these technologies to fulfill their promises?

Vincent Godard – In over 50 years, the Machine Translation market has experienced several technological breakthroughs and SYSTRAN has always been at the forefront of innovation. Recently, Artificial Neural Networks and in-depth studies have made it possible to reach exceptional quality levels, making it possible to increase the use of Machine Translation and to generate strong growth in this market. However, until the Turing test, admit it is comforting to know that it is not yet great. This proves that so far, no one replaces man.

I don’t know if quantum computing, which is in its infancy, will allow us to be even faster, without data loss, but I hope so. It is still impossible to know when quantum computers will be able to do the tasks in minutes that current supercomputers will take thousands of years. It requires a lot of machine learning resources. Unfortunately, the power of computing is still an obstacle to the development of machine translation. Quantum computing really allows us to free ourselves from the power of computing, but like I said above, we’re not there yet.

Leave a Comment