A Streamlined Data Workflow for AI Automation

Track: Multilingual AI | TA5 |   Beginner |
Thursday, November 3, 2022, 9:00am – 9:45am
Held in: Pine/Cedar
Presenter:
Zhenhui Chao - VMware
Host: Tomas Franc

Machine translation (MT) becomes a critical part of the localization process. With all kinds of different data, how do you get the insight of the data, and monitor and optimize the MT quality automatically in your localized content by applying machine learning (ML) techniques? How do you build the quality analytics framework? In this session, we will describe a process starting with collecting the daily operation data, then cleaning up the data, using the data to analyze MT quality, train ML models, deploy the ML services to improve MT quality, and build the analytics framework to get the insight of the MT quality and other data pattern, in an automatic way.

Takeaways: Attendees will learn how to build a data collecting matrix; how to set up a framework to automate the whole process; and how to evaluate the translation quality with visualized reports.