chapter five
5 Advanced Active Learning
This chapter covers
- Understanding different methods for combining Uncertainty Sampling & Diversity Sampling in Active Learning to strategically obtain the optimal samples.
- Implementing machine learning models that predict which unlabeled items will be incorrectly labeled by the model, to identify the most uncertain items to sample for human annotation.
- Implementing machine learning models that predict which unlabeled items are the least like the training data, to identify the most representative items to sample for human annotation.
- Architecting Adaptive Transfer Learning for Representative Sampling and Uncertainty Sampling, to increase the diversity of items sampled in one iteration of Active Learning.
In the last two chapters, you learned how to identify where your model is uncertain (what your model “knows it doesn’t know”), and what is missing from your model (what your model “doesn’t know that it doesn’t know”). In this chapter, you will learn how to combine these into a comprehensive Active Learning strategy.
The use cases in the last two chapters are equally relevant here. For almost any machine learning task, you will want to use both Uncertainty Sampling and Diversity Sampling in combination to sample the optimal items for human review.