"Active Learning" or "Human in the Loop", "Interactive Adaptive Learning", or "Active Learning 101". Want to know what that is?

This session will give a high level description of active learning: what is it, why should we use it?

Active learning is a technique to use when you have not enough labeled data. Since if you have a lot of unlabeled data and want to train a classifier, an option for labelling is to ask a human (or any other information source) - but this will generate costs and might take time since the human is not readily available.

Active learning helps to train a classifier with minimal costs, while reaching the same performance.