Module 5. Training Machine Translation Systems
by Arle Lommel
September 26, 2018
September 26, 2018

Download this research
When language service providers work with advanced MT systems, they need to train them with relevant data to improve quality and relevance to their and their clients’ needs. This module examines how implementers accomplish this vital task. It focuses on the types – and amounts – of data you will need, with an emphasis on translation memory and terminology assets, including what to do when you do not have enough data. It then examines the essential role of data cleaning and preparation in delivering results that will meet quality requirements and the actual process of training the engine. It closes with guidance on special concerns for specific types of systems. This module is intended for use by technology leads and others who will be involved with training and maintaining their MT systems or who will work with MT providers to deliver needed data.

Videos in this series:
Additional modules coming soon:
  • Module 6. Building a Machine Translation-Friendly Supply Chain
  • Module 7. How to Sell Machine Translation Services
  • Module 8. Measuring and Improving Machine Translation Quality
  • Module 9. Advanced Steps to Improve Machine Translation
  • Module 10. Beyond Machine Translation: Linking to Other Technologies

  • Industry Providers
  • Technology Vendors

Pages: 1

Related research
Find similar research by  

Terms of Use | Privacy Statement | Contact Us
Copyright © 2018 Common Sense Advisory, Inc. All Rights Reserved.