Setting Default Labeling Values (Conversational Text)

Hi all. I have a project I’m working on for labeling conversational texts. We have a categorical label “Update” with options “Yes” and “No”. We expect the answer to be “No” most of the time, but it is really important to know when the answer is yes. Is there any way to auto-populate or set the default value of this label to “No” to save labelers time? If not, are there any workarounds people can think of? Thanks!

hey @sjankovic !

Yes, you can auto-populate the value to be “No” to save labelers time.
The way to do this is to upload “Model assisted labeling pre-labels”

The model-assisted labeling (MAL) workflow allows you to import computer-generated predictions (or simply annotations created outside of Labelbox) as pre-labels on an asset.
The imported annotations will be pre-populated in the labeling editor. However, in order to convert the pre-labels to real annotations, a human labeler will still need to open the data row in the editor and submit it. This functionality is designed to speed up human labeling.

Basically, these MAL pre-labels will be pre-populated in the labeling editor, and labelers can overwrite them when labeling.

Here’s a jupyter notebook guide about how to upload them for conversational text.

Let us know if you can’t make it work.

1 Like