The greatest barrier to global trade is not space and time, or even geopolitics. The greatest barrier today is language. Only major corporations have been able to set up mirror offices and global operations networks to facilitate international trade. Human translators, an expensive resource, act as bridges between producers and consumers or business partners who speak different languages. But this is an imperfect solution to a significant challenge. Very few people are fluent, or even competent, in more than a handful of languages, making it difficult for companies to do business in more than a handful of countries.
Much of the world's economy is comprised of small businesses that do not have the resources to overcome dozens of language barriers in countries all over the world. But if small businesses could overcome the language barrier, what would the global market look like? Where would we be if the 'mom and pop' stores that fuel our local economies were able to join the global market?
"There is a wealth of potential in start-ups and entrepreneurial businesses all around the world," says Denis Gachot, CEO of Systran Group. "One of the biggest obstacles to harnessing that potential is the language barrier. Neural Machine Translation (NMT) is working to bring that wall down."
Neural Machine Translation
A natural step in the progression of communication technologies, NMT is a tool that connects people who would otherwise have no means to understand one another. The technology is aimed at translating large volumes of business communications almost instantly and with more effectiveness than human operators.
Unlike preceding translation technologies, NMT builds a neural center of information that can be tuned for more apt translations. The network approach sidesteps the bottleneck often seen in translation technology that hinders the improvement of encoder-decoder systems. This "soft-alignment" is a reflection of our own intuition, so these language translations done by a machine are more human than ever. "This technology is of a caliber that deserves the attention of everyone in the field," says Gachot. "It can translate at near-human levels of accuracy and can translate massive volumes of information exponentially faster than we can operate."
NMT is an end-to-end learning system, so it learns and corrects itself through continued use. Picking up patterns in languages for more accurate translations, NMT systems will continue to improve with time and application, making them the ideal employee.
But part of the beauty of this system is that it is a scalable technology and capable of processing volumes of information exponentially faster than humans are able to. This makes NMT an affordable option for small businesses looking to take their products into more markets. Doors previously closed because of the language barrier are now swinging open. And the world is changing for it.
Companies Using NMT as Plugin
The last year has seen huge leaps in NMT technology. While it remains a tool predominantly for larger businesses, NMT is proving its salt as a revolutionary force in the international market. In September, Google researchers announced their version for this technology, which translates entire sentences instead of just single words, providing a more authentic and relevant translation. Currently, it functions in eight major world languages, able to service some 35 percent of the earth's population.
Already, NMT is being used in clouds and network plugins. Facebook announced in March of last year that they would be using neural networks for their own page translations.
But most exciting in this progression of NMT is its ability to change the way small businesses have been able to contribute to the global market. Google and Systran are racing to roll out NMT in new languages, already delivering dozens of language pairs. The linguistic technology now facilitates communication in 130 different languages, providing real solutions for internal collaboration, online customer support and eDiscovery in multilingual contexts.
What this means practically is that an online store owner in St. Petersburg, Russia is now able to reach a customer in Montevideo, Uruguay and market and discuss her product with reliable translation technology as the go-between.
Small businesses are key to economic strength. Properly equipped with tools and resources, they have the power to grow and shape economies at any scale. Smarter regulations and tax structures are not the only factors at play here. These businesses need to be able to pursue consumers wherever they may be.
According to the U.S. Small Business Administration (SBA), 99.7 percent of all employer firms are small businesses. They have generated 64 percent of new jobs, and paid 44 percent of the total United States private payroll in the last 20 years. The prospect of enabling these small businesses to broaden their customer base and compete in the global market is more than exciting; it is world-changing.
Gachot adds, "For many years, we have tried to deal with language as if it was a barrier for communication - while with neural machine translation, language difference is what makes the richness of communication between different cultures."
NMT looks to be the pivoting step for global trade. With the language wall crumbling, more small businesses will be able to bring their unique, quality-crafted products and services onto the international market scene, which is good for everyone.