Alibaba’s DAMO Academy global research initiative has won five awards at the the Workshop on Machine Translation 2018 (WMT2018).
The company beat contestants from around the world to win in the categories of English to Chinese, English to Russian, Russian to English, English to Turkish and Turkish to English.
The results were announced by the Workshop on Machine Translation last week. The competition, one of the largest machine translation contests, used the US Bilingual Evaluation Understudy (BLEU) international standard to assess the competitors.
WMT has been used for over a decade as a platform for tech companies and academic institutes to showcase their machine learning capabilities. This year’s contestants included teams from Tencent, Microsoft, John Hopkins University and the University of Edinborough.
Alibaba estimates that machine translation is used across its e-commerce market places an average of 750 million times per day.
Improving the accuracy of automatic translations will be key to overcoming the language barrier and allowing merchants to more effectively sell to overseas customers.
“Machine translation is a critical technology for Alibaba’s global business. Alibaba built a world class team in machine translation and natural language processing technologies,” commented Dr Luo Si, leading scientist of natural language processing, machine technologies at Alibaba’s DAMO Academy.
“The team developed the state-of-the-art technologies which empower the business for serving hundreds of millions of customers on Alibaba’s e-commerce platforms and generate huge business values. The results from WMT reflect the excellence of Alibaba’s technologies in the field.”