Overview

In QuickGraph, annotations are categorised into three distinct types: Suggested, Accepted and Agreed Upon. Each annotation type has unique chracteristics and implications. The table below provides a succinct summary of these categories:

Annotation Type	Description
Suggested (Weak)	These annotations have not been verified by an annotator and often result from dataset preannotation or annotation propagation events. In the Annotation View, they appear as semi-opaque entities labels (see Figure 1). The Knowledge Graph displays them as semi-opaque squares.
Accepted (Silver)	These annotations have been verified by an annotator, either through manual application, by accepting a suggestion, or specified in a preannotated dataset (see Figure 1). In the Annotation View, they appear as fully coloured entities labels. The Knowledge Graph represents them as fully coloured circles.
Agreed Upon (Gold)	These annotations represent consensus, achieving majority agreement amongst annotators. For single annotator projects, they are identical to silver annotations. They are prominently featured in the Knowledge Graph and Downloads. They are typically used to train machine learning models.

This tripartite annotation system enhances the quality control process, allows for an iterative and progressive validation of annotations, and facilitates a smoother collaboration among annotators.