CHAIN：通过 Lipschitz 连续性约束归一化增强数据高效 GAN 的泛化能力
CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization
AI总结：
针对数据受限 GAN 中 BN 导致梯度爆炸及判别器过拟合的问题，本文提出 CHAIN 归一化，通过零均值正则化、Lipschitz 约束和自适应特征插值提升稳定性与泛化，并在多个基准上达到 SOTA。
摘要：
生成对抗网络（GAN）显著推动了图像生成的发展，但其性能严重依赖大量训练数据。在数据有限的场景中，GAN 往往面临判别器过拟合和训练不稳定的问题。批归一化（BN）虽以提升泛化能力和训练稳定性而闻名，却很少被用于数据高效 GAN 的判别器。我们的工作针对这一空白，指出 BN 中一个关键缺陷：在中心化和缩放步骤中容易出现梯度爆炸。为解决该问题，我们提出 CHAIN（lipsCHitz continuity constrAIned Normalization，Lipschitz 连续性约束归一化），用零均值正则化替代传统中心化步骤，并在缩放步骤中加入 Lipschitz 连续性约束。CHAIN 还通过在归一化与未归一化特征之间进行自适应插值来增强 GAN 训练，有效避免判别器过拟合。我们的理论分析有力证明，CHAIN 能够降低潜在特征和权重中的梯度，提升 GAN 训练的稳定性与泛化能力。实证结果支持我们的理论。CHAIN 在 CIFAR-10/100、ImageNet、5 个低样本和 7 个高分辨率少样本图像数据集的数据受限场景中取得了最先进结果。代码：https://github.com/MaxwellYaoNi/CHAIN
英文摘要：
Generative Adversarial Networks (GANs) significantly advanced image generation but their performance heavily depends on abundant training data. In scenarios with limited data, GANs often struggle with discriminator overfitting and unstable training. Batch Normalization (BN), despite being known for enhancing generalization and training stability, has rarely been used in the discriminator of Data-Efficient GANs. Our work addresses this gap by identifying a critical flaw in BN: the tendency for gradient explosion during the centering and scaling steps. To tackle this issue, we present CHAIN (lipsCHitz continuity constrAIned Normalization), which replaces the conventional centering step with zero-mean regularization and integrates a Lipschitz continuity constraint in the scaling step. CHAIN further enhances GAN training by adaptively interpolating the normalized and unnormalized features, effectively avoiding discriminator overfitting. Our theoretical analyses firmly establishes CHAIN's effectiveness in reducing gradients in latent features and weights, improving stability and generalization in GAN training. Empirical evidence supports our theory. CHAIN achieves state-of-the-art results in data-limited scenarios on CIFAR-10/100, ImageNet, five low-shot and seven high-resolution few-shot image datasets. Code: https://github.com/MaxwellYaoNi/CHAIN
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
