When Generative Augmentation Hurts: A Benchmark Study of GAN and Diffusion Models for Bias Correction in AI Classification Systems

• URL: http://arxiv.org/abs/2603.16134v1
• Date: Tue, 17 Mar 2026 05:37:17 GMT
• Title: When Generative Augmentation Hurts: A Benchmark Study of GAN and Diffusion Models for Bias Correction in AI Classification Systems
• Abstract summary: Generative models are widely used to compensate for class imbalance in AI training pipelines.<n>FastGAN augmentation does not merely underperform at very low training set sizes but actively increases bias.<n> Stable Diffusion with Low-Rank Adaptation produced the best results overall.
• Score: 0.6875312133832079
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative models are widely used to compensate for class imbalance in AI training pipelines, yet their failure modes under low-data conditions are poorly understood. This paper reports a controlled benchmark comparing three augmentation strategies applied to a fine-grained animal classification task: traditional transforms, FastGAN, and Stable Diffusion 1.5 fine-tuned with Low-Rank Adaptation (LoRA). Using the Oxford-IIIT Pet Dataset with eight artificially underrepresented breeds, we find that FastGAN augmentation does not merely underperform at very low training set sizes but actively increases classifier bias, with a statistically significant large effect across three random seeds (bias gap increase: +20.7%, Cohen's d = +5.03, p = 0.013). The effect size here is large enough to give confidence in the direction of the finding despite the small number of seeds. Feature embedding analysis using t-distributed Stochastic Neighbor Embedding reveals that FastGAN images for severe-minority breeds form tight isolated clusters outside the real image distribution, a pattern consistent with mode collapse. Stable Diffusion with Low-Rank Adaptation produced the best results overall, achieving the highest macro F1 (0.9125 plus or minus 0.0047) and a 13.1% reduction in the bias gap relative to the unaugmented baseline. The data suggest a sample-size boundary somewhere between 20 and 50 training images per class below which GAN augmentation becomes harmful in this setting, though further work across additional domains is needed to establish where that boundary sits more precisely. All experiments run on a consumer-grade GPU with 6 to 8 GB of memory, with no cloud compute required.

Related papers

• Spatial Attention Supervision for Defect Localization: Exploiting Ground-Truth Masks as Training Signal in Diffusion-Augmented Defect Detection [0.0]
  Ground-truth defect masks in industrial datasets are typically reserved for evaluation.<n>This paper repurposes them as spatial supervision signals during training of classification networks.
  arXiv  Detail & Related papers  (2026-09-05T19:20:23Z)
• Magnification-Invariant Image Classification via Domain Generalization and Stable Sparse Embedding Signatures [0.0]
  Magnification shift is a major obstacle to robust histopathology classification, because models trained on one scale often generalize poorly to another.<n>We evaluated this problem on the BreaKHis dataset using a strict patient-disjoint leave-one-magnification-out protocol.
  arXiv  Detail & Related papers  (2026-04-28T16:26:14Z)
• A Heterogeneous Ensemble for Multi-Center COVID-19 Classification from Chest CT Scans [0.39998518782208775]
  We present a heterogeneous ensemble of nine models spanning three inference paradigms.<n>The ensemble achieves an average macro F1 of 0.9280 across four hospital centres, outperforming the best single model.
  arXiv  Detail & Related papers  (2026-03-15T21:34:59Z)
• GRASP: Guided Residual Adapters with Sample-wise Partitioning [10.504309161945065]
  We propose GRASP: Guided Residual Adapters with Sample-wise Partitioning.<n>On the long-tail MIMIC-CXR-LT dataset, GRASP yields superior FID and diversity metrics, especially for rare classes.
  arXiv  Detail & Related papers  (2025-12-01T13:43:17Z)
• DiffusionNFT: Online Diffusion Reinforcement with Forward Process [99.94852379720153]
  Diffusion Negative-aware FineTuning (DiffusionNFT) is a new online RL paradigm that optimize diffusion models directly on the forward process via flow matching.<n>DiffusionNFT is up to $25times$ more efficient than FlowGRPO in head-to-head comparisons, while being CFG-free.
  arXiv  Detail & Related papers  (2025-09-19T16:09:33Z)
• Adapt in the Wild: Test-Time Entropy Minimization with Sharpness and Feature Regularization [85.50560211492898]
  Test-time adaptation (TTA) may fail to improve or even harm the model performance when test data have mixed distribution shifts.<n>This is often a key obstacle preventing existing TTA methods from being deployed in the real world.<n>We propose a sharpness-aware and reliable entropy minimization method, called SAR, for stabilizing TTA from two aspects.
  arXiv  Detail & Related papers  (2025-09-05T10:03:00Z)
• Robust Representation Consistency Model via Contrastive Denoising [83.47584074390842]
  randomized smoothing provides theoretical guarantees for certifying robustness against adversarial perturbations.<n> diffusion models have been successfully employed for randomized smoothing to purify noise-perturbed samples.<n>We reformulate the generative modeling task along the diffusion trajectories in pixel space as a discriminative task in the latent space.
  arXiv  Detail & Related papers  (2025-01-22T18:52:06Z)
• Your Diffusion Model is Secretly a Zero-Shot Classifier [90.40799216880342]
  We show that density estimates from large-scale text-to-image diffusion models can be leveraged to perform zero-shot classification. Our generative approach to classification attains strong results on a variety of benchmarks. Our results are a step toward using generative over discriminative models for downstream tasks.
  arXiv  Detail & Related papers  (2023-03-28T17:59:56Z)
• Compound Batch Normalization for Long-tailed Image Classification [77.42829178064807]
  We propose a compound batch normalization method based on a Gaussian mixture. It can model the feature space more comprehensively and reduce the dominance of head classes. The proposed method outperforms existing methods on long-tailed image classification.
  arXiv  Detail & Related papers  (2022-12-02T07:31:39Z)
• Adaptive DropBlock Enhanced Generative Adversarial Networks for Hyperspectral Image Classification [36.679303770326264]
  We propose an Adaptive DropBlock-enhanced Generative Adversarial Networks (ADGAN) for hyperspectral image (HSI) classification. The discriminator in GAN always contradicts itself and tries to associate fake labels to the minority-class samples, and thus impair the classification performance. Experimental results on three HSI datasets demonstrated that the proposed ADGAN achieved superior performance over state-of-the-art GAN-based methods.
  arXiv  Detail & Related papers  (2022-01-22T01:43:59Z)
• Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection [85.53263670166304]
  One-stage detector basically formulates object detection as dense classification and localization. Recent trend for one-stage detectors is to introduce an individual prediction branch to estimate the quality of localization. This paper delves into the representations of the above three fundamental elements: quality estimation, classification and localization.
  arXiv  Detail & Related papers  (2020-06-08T07:24:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.