Overview
In active learning, the model identifies which data points would be most beneficial to learn from and asks for labels for those specific points. This is much more efficient than labeling data at random.
Strategies
- Uncertainty Sampling: Querying the points the model is least certain about.
- Query-by-Committee: Using multiple models and querying points where they disagree.
- Expected Model Change: Querying points that would most change the model's current state.
Goal
To achieve high accuracy with as few labeled examples as possible.