Hover Example Apps

(works out-of-the-box on Binder)

This repository demonstrates the usage of hover to understand and supervise a machine learning task.

Simple Annotator

• use the lasso/tap/poly tool to select data points
• assign your labels and hit "Apply"
• in a real use case on your computer, "Export" will save a DataFrame
• take advantage of the text/regex search box

Linked Annotator

• on top of the Simple Annotator, showes another plot which is focused on search
• the search boxes are independent across plots, minimizing interference in between
• 💡 the selections in the plots are synchronized. You can select in one and label in the other!

Active Learning

• compared with the Linked Annotator, replaces the left plot with the predictions of a model in the loop
• 💡 inspect data points based on their prediction confidence and locations!

Snorkel Annotator

• compared with the Linked Annotator, replaces the left plot with Snorkel labeling function outputs
• in a real use case, you can check the labeling functions and your annotations against each other.
• 💡 one could also exploit labeling functions as sophisticated "filters" to help find interesting data points!