Learning with not Enough Data Part 2: Active Learning
Lilian Weng
Active learning is a machine learning technique that selects which unlabeled samples to label within a fixed budget to maximize model performance improvement, particularly useful when labeling is costly. Key approaches include uncertainty sampling (using prediction confidence or model ensembles), diversity sampling (selecting representative samples), and expected model change, with MC dropout being an economical method for uncertainty estimation in deep neural networks. The method enables more efficient model training by intelligently choosing which samples to label rather than labeling randomly.
Why it matters
This is part 2 of what to do when facing a limited amount of labeled data for supervised learning tasks. This time we will get some amount of human labeling work involved, but within a budget limit, and therefore we need to be smart when selecting which samples to label.
