Gradient-Discrepancy Acquisition for Pool-Based Active Learning

TL;DR

This paper introduces a new gradient-based acquisition criterion for pool-based active learning, justified theoretically and validated empirically, enhancing data point selection strategies.

Contribution

It proposes a novel gradient discrepancy criterion for active learning, integrating theoretical insights with practical effectiveness.

Findings

The gradient discrepancy criterion improves data selection in active learning.

The method is effective across different active learning strategies.

Theoretical justification supports its use as a reliable acquisition measure.

Abstract

The effectiveness of active learning hinges on the choice of the acquisition criterion by which a learning algorithm selects potentially informative data points whose label is subsequently queried. This paper proposes a novel gradient-based acquisition criterion, derived from a generalization bound introduced by Luo et al. (2022). This criterion can be applied in lieu of uncertainty measures in uncertainty sampling, or incorporated into diversity-based methods that consider the spread of sampled points in addition to the uncertainty of their labels. We provide a theoretical justification of the proposed acquisition criterion, and demonstrate its effectiveness in an empirical evaluation.