Learning-State-Aware Dynamic Generative Data Augmentation on Small-Scale Datasets

2026-08-19

AI Summary

本文提出学习状态感知的动态生成数据增强方法LSADA，通过分析样本损失及下降率确定增强强度，并采用解耦增强与融合策略提高小规模数据集分类性能。

Abstract

Small-scale image classification is often limited by the scarcity of training data. Generative data augmentation (GDA) based on pretrained generative models has emerged as an effective solution. However, existing methods rely on task-agnostic augmentation strategies that overlook downstream model needs. Although recent dynamic GDA methods incorporate model feedback to guide augmentation, they still struggle to reliably determine sample-specific augmentation strengths and adapt augmentation strategies to different image regions while balancing image diversity and class semantics. To address these issues, we propose learning-state-aware dynamic generative data augmentation (LSADA). Specifically, LSADA constructs a learning state for each sample based on its current loss and loss-decrease rate, which is then mapped to a sample-specific augmentation strength. Furthermore, LSADA introduces a decoupled data augmentation and diffusion fusion strategy that applies strength-controlled transformations to class-relevant regions and generates diverse class-irrelevant regions, progressively fusing them to improve image diversity while preserving class semantics. Experiments on nine public datasets show that LSADA outperforms the existing SOTA dynamic GDA method by an average of 4.5% on six natural image datasets and 2.5% on three medical image datasets.

Problem

• Small-Scale Datasets
• Generative Data Augmentation
• Dynamic GDA
• Sample-Specific Augmentation Strength
• Image Diversity

Innovation

• learning-state-aware
• dynamic generative data augmentation
• loss-decrease rate
• decoupled data augmentation and diffusion fusion