5 Advanced 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. You will also learn how to use transfer learning to adapt your models to predict which items to sample.

5.1    Combining Uncertainty Sampling and Diversity Sampling

This section will explore ways to combine all the Active Learning techniques that you have learned up to this point so that you can effectively leverage them for your particular use cases. You will also learn one new Active Learning strategy, Expected Error Reduction, that combines principles of Uncertainty Sampling and Diversity Sampling.

Recall from Chapter 1 that an ideal strategy for Active Learning tries to sample items that are near the decision boundary but are distant from each other, as repeated here in Figure 5.1:

5.1.1   Least Confidence Sampling with Clustering-based Sampling

5.1.2   Uncertainty Sampling with Model-based Outliers

5.1.3   Uncertainty Sampling with Model-based Outliers and Clustering

5.1.4   Representative Sampling Cluster-based Sampling

5.1.5   Sampling from the Highest Entropy Cluster

5.1.6   Other Combinations of Active Learning Strategies

5.1.7   Combining Active Learning Scores

5.1.8   Expected Error Reduction Sampling

5.2    Active Transfer Learning for Uncertainty Sampling

5.2.1   Making your model predict its own errors

5.2.2   Implementing Active Transfer Learning

5.2.3   Active Transfer Learning with more layers

5.2.4   The pros and cons of Active Transfer Learning

5.3    Applying Active Transfer Learning to Representative Sampling

5.3.1   Making your model predict what it doesn’t know

5.3.2   Active Transfer Learning for Adaptive Representative Sampling

5.3.3   The pros and cons of Active Transfer Learning for Representative Sampling