Contents
This tutorial on machine translation quality prediction and automatic post-editing is based on what we’ve learned making “quality estimation” research work in the real world.
High-volume, high-quality translation buyers, from Germany's engineering giants to global marketplaces for luxury fashion, are already translating up to 10x more efficiently - at the same human quality - by using private custom LLMs to automatically edit and verify millions of automatically generated translations, right inside their translation management systems.
More and more translation teams are wondering how to get started, or how to catch up.
Adam, one of the co-founders of ModelFront, will hold a practical, accessible hands-on tutorial on making these technologies work in real world translation workflows.
He was previously a software engineer at Google Translate within Google Research, and founded the Machine Translate Foundation.
Takeaways
A practical overview for making machine translation quality prediction and automatic post-editing work in the real world:
-
What it’s for
-
How it works
-
When it’s a fit
-
How to choose
-
What it changes
Prior knowledge
You'll get the most out of this tutorial if you know the basics of traditional fully manual human post-editing of machine translation.