Start annotating manually. At least 5 or 10 annotations per label.
The system will observe what is happening and after a while start to train a first model using your annotation so far in order to calculate “suggestions”.
You may want to continue even after the first appearance of the blue “pop up“ anounces that suggestions have been computed (see What is a suggestion?).
After a decent amount of annotations (at least 10 for the simplest scenario),
- Go to the Suggestions view as below:
- Accept/reject/correct the suggested annotations (green check, red cross…)
- Make sure your segment is properly annotated: no false annotations, no missing annotations, no inconsistencies
- Filter the list of suggestions you want to work on
- Validate the segment (or document): It will be added to the dataset with its annotation and be used in subsequent training runs
- The suggestion engine is updated after few validations.
- Sort suggestions according to their confidence level score.
- Increase the context of the segment if necessary
- Move to a document view in the suggestion user interface. Entities validation is only on the segment. The rest of the text is greyed.
Let’s move on to the next step: