Neural machine translation is the latest translation technology with the potential to overcome many of the weaknesses of conventional phrase-based translation systems, according to KIPO. Currently, Google, Kakao and Naver are providing artificial intelligence translation services using the neural machine translation technologies.
The government agency said it agreed to learn the neural machine translation technologies from the WIPO at the two-day information cooperation meeting held in Geneva from Tuesday.
The WIPO developed neural machine translation in September, 2016 and has since provided translation services in 10 languages -- Arabic, Chinese, English, French, German, Japanese, Korean, Portuguese, Russian and Spanish.
“We have reviewed the adoption of the neural machine translation over the years and we found translation quality is much improved when artificial neuron technology is adopted for machine translation,” said a KIPO spokesperson.
He added the WIPO’s neuron machine translation is especially customized for patent translations as it renders highly technical patent documents into a second language in a style and syntax that more closely mirrors common usage.
The agency will provide the machine translation using the artificial neuron technology through its verification system for patent information for foreign users, who inquire about Korea’s patent screening information.
“We will continue to establish helpful patent administration by discovering additional cooperation projects related to artificial intelligence and information communication technology,” said Moon Sam-seob, a chief of KIPO’s information support division.
The global automated translation market is predicted to grow from $600 million in 2012 to $6.9 billion by 2019, according to the US-based WinterGreen Research.
By Shin Ji-hye (firstname.lastname@example.org)