Alibaba launches AI e-commerce translation tool

25/05/2018

The former supervisor of Facebook’s AI translation team has joined Alibaba to launch the first-ever e-commerce translation tool utilizing artificial intelligence. 

 

 

Alibaba touts global talent

Dr. Fei Huang, the former supervisor of Facebook’s artificial intelligence translation team, has joined Alibaba’s DAMO Academy.

Dr. Huang is preceded by recent hires Dr. Boxing Chen and Dr. Xiaozhong Liu, who also joined the neuro-linguistic programming (NLP) team at the DAMO Machine Intelligence Laboratory.

Dr. Chen has been engaged in machine translation studies at the National Research Council of Canada, and Dr. Xiaozhong Liu has worked in information retrieval (IR) at Yahoo.

 

 

Communication is one of the biggest obstacle in cross-border deals. At Alibaba International Station, 70% of buyers communicate in English, while 30% communicate in other languages.

 

 

Luo Si, head of the NLP team at Alibaba’s DAMO Academy, said that nearly 100 NLP members are currently located in China, the United States and Singapore. More than 30% of the personnel have doctoral degrees, and the team is still recruiting industry elites.

After joining Alibaba, Dr. Huang will promote the real-time translation application in for New Retail, cross-border e-commerce, video and social media.

Communication is one of the biggest obstacle in cross-border deals. At Alibaba International Station, 70% of buyers communicate in English, while 30% communicate in other languages. The seller-side survey data shows that about 96% of sellers are unable to communicate in other languages.

The first e-commerce real-time AI translation tool

On May 18, Alibaba formally launched its real-time translation function in cross-border e-commerce, helping to bridge the language barrier between buyers and sellers. It is reportedly the world’s first real-time translation AI tool created for the e-commerce industry.

Like humans, the app can automatically correct errors based on context and grammar.

The translation system uses the latest neural network model in the industry and integrates many of DAMO Academy’s AI techniques, including machine translation, semantic understanding, automatic language recognition and speech recognition.

The system currently supports four languages: Chinese, English, Russian and Spanish. It is expected that in the second half of the year, translation services will be expanded to include Thai, Arabic, Vietnamese and Turkish.

 

 

Like humans, the app can automatically correct errors based on context and grammar.

 

 

According to public data on the internet, the total number of Amazon’s commodities were 390 million in 2017. But Amazon adopted a nationally operated strategy, so the number of goods that need to be translated is far less than this.

The total amount of goods that Alibaba currently needs to translate exceeds 400 million. With such a large amount of multilingual e-commerce products to handle, Alibaba’s  AI translation system is put to work, translating 750 million times a day.

Alibaba also recently announced it will launch real-time voice and live-video translation video to automatically generate bilingual subtitles. That technology would allow cross-border merchants to directly communicate with one another in their preferred language via video.

Leave a Reply

Your email address will not be published. Required fields are marked *