REAL: A Representative Error-Driven Approach for Active Learning

TL;DR

REAL introduces a novel active learning method that focuses on selecting representative pseudo errors based on error density, leading to improved model accuracy and F1 scores in text classification tasks.

Contribution

It proposes a new error-driven sampling approach that considers neighborhood error density, outperforming existing methods in active learning for text classification.

Findings

Outperforms baselines in accuracy and F1-macro scores

Selects representative pseudo errors matching ground-truth error distribution

Effective across various hyperparameter settings

Abstract

Given a limited labeling budget, active learning (AL) aims to sample the most informative instances from an unlabeled pool to acquire labels for subsequent model training. To achieve this, AL typically measures the informativeness of unlabeled instances based on uncertainty and diversity. However, it does not consider erroneous instances with their neighborhood error density, which have great potential to improve the model performance. To address this limitation, we propose $REAL$, a novel approach to select data instances with $\underline{R}$epresentative $\underline{E}$rrors for $\underline{A}$ctive $\underline{L}$earning. It identifies minority predictions as \emph{pseudo errors} within a cluster and allocates an adaptive sampling budget for the cluster based on estimated error density. Extensive experiments on five text classification datasets demonstrate that $REAL$ consistently outperforms all best-performing baselines regarding accuracy and F1-macro scores across a wide range of hyperparameter settings. Our analysis also shows that $REAL$ selects the most representative pseudo errors that match the distribution of ground-truth errors along the decision boundary. Our code is publicly available at https://github.com/withchencheng/ECML_PKDD_23_Real.