Minimum-Margin Active Learning
This is an incremental improvement for active learning practitioners, enhancing sample efficiency in supervised learning tasks.
The paper tackles the problem of selecting which examples to label in batch active learning by proposing a min-margin method that uses bootstrapped models to choose candidates based on minimum margin, showing it outperforms other methods, especially with larger batch sizes.
We present a new active sampling method we call min-margin which trains multiple learners on bootstrap samples and then chooses the examples to label based on the candidates' minimum margin amongst the bootstrapped models. This extends standard margin sampling in a way that increases its diversity in a supervised manner as it arises from the model uncertainty. We focus on the one-shot batch active learning setting, and show theoretically and through extensive experiments on a broad set of problems that min-margin outperforms other methods, particularly as batch size grows.