Tutorials
Learning-State-Aware Dynamic Generative Data Augmentation on Small-Scale Datasets
arXiv:2608.18907v1 Announce Type: cross Abstract: Small-scale image classification is often limited by the scarcity of training data. Generative data augmentation (GDA) based on pretrained generative
arXiv:2608.18907v1 Announce Type: cross 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.
Related
- Scaling few-shot spoken word classification with generative meta-continual learning
- The two clocks and the innovation window: When and how generative models learn rules
- Feasible and Novel Synthetic Population Generation with Tabular and Sequential Travel Attributes
Source: arXiv cs.AI | 2026-08-20