While machine translations continues to advance, it is never expected to be more than an aid for human translators conducting IP translations, according to an article Jeremy Coombs wrote for IPFrontline.com.
In his article, Coombs outlines three technologies that are simplifying IP translations, especially patents: terminology management, translation memory and machine-assisted technologies. All play a critical role, but need to be used with human translators, who are more intimately familiar with the language, scientific and cultural nuances of a given language.
Terminology management technology “is critical, since brands, technical terms and trade terms can rarely be translated literally, and each translator will make different interpretations and usually end up with different translations," Coombs added.
Translation memory technology lets translators leverage previous translations in new projects. While this speeds the process and helps ensure accuracy, “it is still critical that humans review the final product because the context of the translation must still be determined.”
As for machine-assisted technologies, these train servers to contextually match terms for an initial translation, which must then be edited by a human translator. In addition to assisting human translators, they can also be used in e-discovery or prior art searches. But patent translations are far too important to leave entirely to a machine.
Before deciding on your IP translation provider, ensure they are using the most advanced translation technologies so you’ll benefit from fewer errors, lower costs and subsequently, faster time to grant.
Which other technologies are your service providers using?