The Republic of Korea has become the first adopter of a new artificial intelligence-based translation tool for patent documents created by the World Intellectual Property Organization (WIPO). The tool WIPO Translate uses cutting-edge neural machine translation technology to render highly technical patent documents into a second language in a style and syntax that more closely mirrors common usage.

MoU signed
The Korean Intellectual Property Office (KIPO) and WIPO signed a memorandum of understanding on the sidelines of the May 23-25 meeting of Intellectual Property offices on ICT Strategies and Artificial Intelligence for IP Administration. Under the agreement, KIPO will begin integrating WIPO Translate into its own patent filing and examination processes.  This makes the Republic of Korea the first of WIPO’s 191 member states to implement WIPO Translate under a program designed to help IP offices around the world use the latest technologies. KIPO’s adoption of WIPO Translate underlines the importance of this industry leading tool, which promotes the widespread dissemination of the knowledge contained in highly technical patent documents,’ said WIPO Director General Francis Gurry. Neural machine translation produces more natural word order, and particularly effective in so-called distant language pairs like, Korean-English, Japanese-English or Chinese-English.

Out-performing
WIPO says the tool out-performs patent-translation tools built on previous technologies as well as other web-based products also using artificial intelligence. It powers PATENTSCOPE, a 70 million record strong database used for research by inventors before filing international patent applications around the world via WIPO’s Patent Cooperation Treaty (PCT). WIPO has “trained” the new technology to translate all patent documents in one of the official languages of the PCT (Arabic, Chinese, French, German, Japanese, Korean, Portuguese, Russian and Spanish) into English and vice-versa. To develop WIPO Translate, WIPO created its own software, based on open-source software and libraries and capitalized on in-house expertise in handling large datasets. Readers can try out the tool here.