Unbabel’s workflow: crowdsourcing model and scalable quality management processes thumbnail
slide-image
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Unbabel’s workflow: crowdsourcing model and scalable quality management processes

Published on Apr 08, 2019101 Views

Related categories

Chapter list

Unbabel’s workflow: Crowdsourcing model and scalable quality management00:00
Quality in a AI + Crowd pipeline00:20
Unbabel Pipeline01:17
Distributed Pipeline02:08
Dynamic Factors on Quality02:36
Editors Pool05:29
Editor’s on-boarding06:36
Crowd Assessment07:17
Crowd Assessment08:25
Evaluation Error Criteria09:02
How Editors are Evaluated09:30
Editors Profiling09:49
Editor Assignment09:53
Editors Pool10:09
Error typologies10:12
MQM (QT21)10:41
MQM Core11:10
MQM (QT21)11:17
Annotation at Unbabel11:55
ErrorTypology for Annotation12:34
Deep Annotations12:39
Annotation at Unbabel12:42
Smartcheck 12:52
Smartcheck: Learn13:02
Pos-Editing Interfaces13:03
Thank you!13:06