Wipo (World Intellectual Property Organization) has developed a translation tool for patent documents. Wipo Translate incorporates a 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, outperforming other translation tools built on previous technologies. Wipo has initially “trained” the new technology to translate Chinese, Japanese and Korean patent documents into English. As indicated by Wipo, patent applications in those languages accounted for some 55% of worldwide filings in 2014. “One of the aims of the patent system is to make technology available. Language is a barrier to the universal achievement of that aim. This breakthrough for Wipo Translate means that a vast, and ever-increasing, trove of patent documents will soon be more easily accessible to innovators who search these records for inspiration or technical know-how – said Wipo director general Francis Gurry -. As part of a global trend, patent applications are increasingly being filed in East Asian languages, particularly in Chinese, and Wipo Translate helps ensure that state-of-the-art knowledge created in these languages is shared as widely and rapidly as possible”. The high level of accuracy of the Chinese-English translation is the result of the training of the neural machine translation tool, which compared 60 million sentences from Chinese patent documents provided to Wipo’s Patentscope database by the State Intellectual Property Office of the People’s Republic of China with their translations as filed at the United States Patent and Trademark Office. Wipo plans to extend the neural machine translation service to French-language patent applications, with other languages to follow. The Patentscope database integrates with other translation engines freely available on the internet and continues to use existing statistical-based translation technology for languages where it performs well. Neural machine translation is an emerging technology. It is based on huge neural network models that “learn” from previously translated sentences. The specificity of neural machine translation (compared to previous “phrase based” statistical methods) is that it produces more natural word order.
