At seven o'clock in the morning, in Shandong Jincheng Carpet Co., Ltd, a transportation truck from Qingdao Customs is nervously loading boxes. "This is sent to Australia, yesterday went three cars, today went one, a total of four cabinets." The person in charge of the company, Gu Guoming, told reporters happily. After seeing the trucks leave, Gu Guoming could not rest and non-stop to the workshop to check the production of another batch of export goods. For him, such a busy life has become a daily routine.
Gu Guoming's English is not good, small language is not, but riding on the cross-border e-commerce "east wind", Gu Guoming's small company can now reach 24 million yuan of annual foreign trade sales, and lead the village collective income of 7 million yuan a year. In the context of the collective downturn of various industries caused by the epidemic, such achievements are bright enough.
Not coincidentally, known as the "hometown of knitting" in Guangxi Bobai, there are more than 400 small knitting enterprises with the help of cross-border e-commerce to open up a new business situation, the past two years each year, a total of more than 1 billion yuan of foreign trade revenue.
Undeniably, cross-border business is becoming a life-saving straw for countless small and medium-sized enterprises under the epidemic. Surprisingly, with the evolving platform translation technology, "language" is no longer the number one roadblock in the cross-border transaction chain, millions of small and medium-sized enterprises like Gu Guoming can communicate seamlessly with overseas buyers every day.
"Language" is no longer a roadblock in foreign trade business.
At the beginning of 2020, the outbreak of the Xin Guan epidemic forced countless SMEs to press the pause button on their business. Fortunately, a few months later, due to the effective control of the epidemic, China resumed production order earlier than other countries, and a large number of overseas orders flooded into the country. Many business owners who had not previously been involved in foreign trade also looked to overseas markets.
However, cross-border business is not good, the first thing they face is the problem of cross-language communication with foreign customers.
Research shows that in Alibaba International Station, 70% of buyers communicate in English, the remaining 30% use small languages, on the other hand, about 96% of sellers can not do anything about small languages, and many sellers are not even proficient in English. Therefore, sellers like Guoming need to use the automatic translation function of the platform to release goods and communicate with customers smoothly.
In the face of cross-border e-commerce translation of large volume, information fragmentation, multilingual and other challenges, major platforms have also been in the breakthroughs, and strive to build a good cross-border e-commerce bridge between buyers and sellers.
Take the Ali platform currently used by Guoming as an example, its machine translation provides multi-language real-time translation capability in the entire cross-border e-commerce chain, effectively solving the language problem of cross-border e-commerce platforms in various links. For example, at the stage of product release, sellers can click "one-click release" after entering Chinese, and the platform will automatically translate the information into the corresponding foreign language and release it on the international platform. When communicating with buyers, regardless of the language of communication, the software can always be translated into Chinese in real time, so that sellers can hardly feel the existence of the translation AI behind the platform.
Currently, Ali Translation covers 214 languages and provides hundreds of millions of words of translation services for 2 million domestic small and medium-sized merchants every day, helping them sell their products from all corners of China to the world.
After opening up the language communication barriers of cross-border e-commerce, the village has become a "digital sea".
Relying on platform translation technology to broaden the boundary of cross-border e-commerce
Ali translation technology development process, is China's cross-border e-commerce platform translation technology continues to progress epitomized. In the case of technology originally lagging behind overseas, the team step by step from scratch to build a leading translation technology system, so that the algorithm gradually from the "can be used" across to "good", for small and medium-sized enterprises to participate in the cross-border export trade to smooth out the language barrier.
t is understood that since 2009, Ali began to use the open source translation system to provide translation services for merchants in the field of e-commerce, but at that time, the translation was more mechanical, and the effect was not ideal. 2013, in order to improve the accuracy of the translation, Ali spent three years to set up an independent technical team, began to build a statistically based machine translation technology, and the introduction of artificial translation services as a supplement. By 2016, with the development of artificial intelligence, Ali built a translation technology system based on neural networks and began to provide automatic translation services in multiple languages and fields, and in 2018, Ali participated in the WMT 2018 World Machine Translation Competition for the first time to win the first place in five language directions.
Since then, Ali has gradually opened up its translation capabilities to serve a large number of software developers and small and medium-sized enterprises in the cross-border industry, adding fuel to the development of the entire cross-border e-commerce industry, while building more general multilingual translation capabilities.2021, Alibaba was awarded the Ministry of Industry and Information Technology's Artificial Intelligence Industry Innovation "Intelligent Translation System", and the "field" was unveiled. In 2021, Alibaba was awarded the Ministry of Industry and Information Technology's Artificial Intelligence Industry Innovation "Intelligent Translation System" and "Unveiled and Commanded" Winning Unit.
Behind this step-by-step progression, it is inseparable from the Ali natural language processing technology team's unremitting refinement of technology.
Yang Baosong is an engineer in the Language Technology Lab of Ali Dharma Institute, and his group specializes in supporting the research and development of translation algorithms for cross-border e-commerce platforms. Every week, he and his colleagues collect feedback from buyers and sellers on the platform, and every quarter the team visits merchants on the ground to gain a deeper understanding of user needs, so as to optimize and upgrade the translation platform. For example, in the past two years, in order to further improve the "authenticity" of translation on the basis of accuracy, the team pioneered the "creative translation technology" in the industry, i.e., the creative translation technology, which enables the system to automatically make corresponding adjustments according to the language habits and cultures of various countries, such as reordering the product information and translating it into Italian. The system can automatically make corresponding adjustments according to the language habits and cultures of different countries, such as reordering, paraphrasing, merging, expanding and so on.
"In China, people like to call the product as Baobei, but if we directly translate it as 'baby' to foreign consumers, people will find it very strange, at this time, our translation technology can automatically adjust to something like ' product'." Yang Baosong said.
Many domestic merchants are accustomed to piling up a large number of keywords in the title, which often brings reading obstacles to overseas buyers, and this problem can be improved after applying the translation technology. Yang Baosong introduced, "different countries buyers prefer different reading order, such as South Korean buyers like to see the brand and core selling points first, Thai buyers like to see the discounts first, our translation system can automatically according to the habits of overseas buyers on the title to streamline and refine, and adjust the order of information."
The difficulty of this new technology development is quite high, because there is no mature technology at home and abroad to learn from, the team can only feel the stones to cross the river, early through manual one by one to summarize the commodities of different countries, and cumulative study of hundreds of thousands of commodities. At a later stage, through the accumulation of technology, the system finally realized the automatic learning of expression styles of different places.
Market demand is the best incubator for advanced technology. Right now, cross-border e-commerce will further explode, some data show that by 2024, China's cross-border export e-commerce transactions are expected to exceed 2.95 trillion yuan. It is foreseeable that the booming cross-border e-commerce will also force the platform translation technology to usher in greater breakthroughs.
Source: China Daily
English 
简体中文