Diversity Enhanced Active Learning with Strictly Proper Scoring Rules
This work addresses improving active learning efficiency for text classification, representing an incremental advancement in acquisition function design.
The paper tackled the problem of designing acquisition functions for active learning in text classification by proposing BEMPS, which estimates increases in strictly proper scores, and a complementary batch algorithm for diversity. The results showed that BEMPS with mean square error and log probability consistently outperformed other tested methods.
We study acquisition functions for active learning (AL) for text classification. The Expected Loss Reduction (ELR) method focuses on a Bayesian estimate of the reduction in classification error, recently updated with Mean Objective Cost of Uncertainty (MOCU). We convert the ELR framework to estimate the increase in (strictly proper) scores like log probability or negative mean square error, which we call Bayesian Estimate of Mean Proper Scores (BEMPS). We also prove convergence results borrowing techniques used with MOCU. In order to allow better experimentation with the new acquisition functions, we develop a complementary batch AL algorithm, which encourages diversity in the vector of expected changes in scores for unlabelled data. To allow high performance text classifiers, we combine ensembling and dynamic validation set construction on pretrained language models. Extensive experimental evaluation then explores how these different acquisition functions perform. The results show that the use of mean square error and log probability with BEMPS yields robust acquisition functions, which consistently outperform the others tested.