ALVIN: Active Learning Via INterpolation

TL;DR

ALVIN introduces a novel active learning method that uses intra-class interpolations to select informative, under-represented examples, improving model robustness and generalization across diverse datasets.

Contribution

The paper proposes ALVIN, a new active learning approach that creates artificial anchors through interpolation to better identify informative examples and mitigate shortcut learning.

Findings

ALVIN outperforms existing active learning methods on six datasets.

It improves both in-distribution and out-of-distribution generalization.

Experimental results show enhanced model robustness against shortcuts.

Abstract

Active Learning aims to minimize annotation effort by selecting the most useful instances from a pool of unlabeled data. However, typical active learning methods overlook the presence of distinct example groups within a class, whose prevalence may vary, e.g., in occupation classification datasets certain demographics are disproportionately represented in specific classes. This oversight causes models to rely on shortcuts for predictions, i.e., spurious correlations between input attributes and labels occurring in well-represented groups. To address this issue, we propose Active Learning Via INterpolation (ALVIN), which conducts intra-class interpolations between examples from under-represented and well-represented groups to create anchors, i.e., artificial points situated between the example groups in the representation space. By selecting instances close to the anchors for annotation, ALVIN identifies informative examples exposing the model to regions of the representation space that counteract the influence of shortcuts. Crucially, since the model considers these examples to be of high certainty, they are likely to be ignored by typical active learning methods. Experimental results on six datasets encompassing sentiment analysis, natural language inference, and paraphrase detection demonstrate that ALVIN outperforms state-of-the-art active learning methods in both in-distribution and out-of-distribution generalization.