Integrating Custom-trained Neural Machine Translation into a Continuous Localization Flow in Enterprises

Track: Automation | AU9 |   Intermediate |
Thursday, July 30, 2020, 3:00am – 3:30am
Held in: Stream 1
Alexander Gigga - Lengoo 
Giorgio Mattiuzzo - NI

Switching from a decentrally organized CAT-assisted localization workflow to a centralized custom neural machine translation (NMT) and post-editing workflow in an enterprise can be disruptive. In this session we will present a case study of the cooperation between lengoo and National Instruments. We will share the challenges and possibilities for organizations to integrate custom NMT into their localization flows and showcase the tremendous potential of the technology in terms of language quality, time-to-market and cost-efficiency. We will also focus on the responsibilities of both sides concerning data quality, engine training and quality assurance.

Takeaways: Attendees will learn about the challenges of the implementation of custom NMT from a vendor’s and a buyer’s perspective, the cost and time advantages to the localization flow and the organizational impact it can have on an internal localization unit.