Once the documents containing foreign languages are identified, and the critical foreign language(s) are selected, ILS employs machine translation to enable English speaking reviewers to quickly identify and triage relevant documents without the need to first incur costly and time-consuming human translations. But not all machine translation systems are created equal.
ILS created an analytics driven machine translation workhorse and hired linguists to teach it particular foreign languages with specific focus on case specific subject matter. For example, in the Toyota Unintended Acceleration Cases, we had to program for many nuances of Japanese language like differentiating between Tundra as frozen ground and Tundra as a pickup truck.
We’ve developed workflows, algorithms, and processes to find gaps in productions, to identify missing custodians, to uncover the hidden truths that Plaintiff counsel knows in their gut must exist but which often require ILS’ special technology and acumen to find. We understand the importance of utilizing native speakers to program the machine so that we use the construction of native language, including idioms, local expressions and slang to find instances of the concepts being sought. We test our results to fine tune them so that we can rely on the machine to find the kernels of evidence we’re looking for amongst the massive document dumps we’ve received.