Beyond Loss Values: Robust Dynamic Pruning via Loss Trajectory Alignment

TL;DR

AlignPrune introduces a loss-trajectory-based criterion called Dynamic Alignment Score to improve the robustness of dynamic data pruning methods under noisy-label conditions, enhancing accuracy without altering existing frameworks.

Contribution

The paper proposes AlignPrune, a noise-robust module utilizing loss trajectory alignment to better identify noisy samples during pruning, outperforming existing methods.

Findings

Boosts accuracy by up to 6.3% over baselines.

Effective across various noise types and pruning ratios.

Seamlessly integrates into existing pruning frameworks.

Abstract

Existing dynamic data pruning methods often fail under noisy-label settings, as they typically rely on per-sample loss as the ranking criterion. This could mistakenly lead to preserving noisy samples due to their high loss values, resulting in significant performance drop. To address this, we propose AlignPrune, a noise-robust module designed to enhance the reliability of dynamic pruning under label noise. Specifically, AlignPrune introduces the Dynamic Alignment Score (DAS), which is a loss-trajectory-based criterion that enables more accurate identification of noisy samples, thereby improving pruning effectiveness. As a simple yet effective plug-and-play module, AlignPrune can be seamlessly integrated into state-of-the-art dynamic pruning frameworks, consistently outperforming them without modifying either the model architecture or the training pipeline. Extensive experiments on five widely-used benchmarks across various noise types and pruning ratios demonstrate the effectiveness of AlignPrune, boosting accuracy by up to 6.3\% over state-of-the-art baselines. Our results offer a generalizable solution for pruning under noisy data, encouraging further exploration of learning in real-world scenarios. Code is available at: https://github.com/leonqin430/AlignPrune.