对抗分数蒸馏：当分数蒸馏遇见 GAN
Adversarial Score Distillation: When score distillation meets GAN
AI总结：
本文用 WGAN 范式揭示现有分数蒸馏因判别器次优或优化不完整而对 CFG 尺度敏感，提出采用可优化判别器和完整目标的 ASD，并在 2D 蒸馏、文本生成 3D 与图像编辑任务中验证其效果。
摘要：
现有的分数蒸馏方法对无分类器引导（CFG）尺度很敏感：在较小的 CFG 尺度下表现为过度平滑或不稳定，而在较大尺度下则会过度饱和。为解释和分析这些问题，我们重新审视分数蒸馏采样（SDS）的推导过程，并利用 Wasserstein 生成对抗网络（WGAN）范式来解读现有分数蒸馏。借助 WGAN 范式，我们发现现有分数蒸馏要么采用固定的次优判别器，要么进行不完整的判别器优化，从而导致尺度敏感问题。我们提出对抗分数蒸馏（ASD），它维护一个可优化的判别器，并使用完整的优化目标对其进行更新。实验表明，所提出的 ASD 在 2D 蒸馏和文本生成 3D 任务中相较现有方法表现出色。此外，为探索我们 WGAN 范式的泛化能力，我们将 ASD 扩展到图像编辑任务，并取得了具有竞争力的结果。项目页面和代码位于 https://github.com/2y7c3/ASD。
英文摘要：
Existing score distillation methods are sensitive to classifier-free guidance (CFG) scale: manifested as over-smoothness or instability at small CFG scales, while over-saturation at large ones. To explain and analyze these issues, we revisit the derivation of Score Distillation Sampling (SDS) and decipher existing score distillation with the Wasserstein Generative Adversarial Network (WGAN) paradigm. With the WGAN paradigm, we find that existing score distillation either employs a fixed sub-optimal discriminator or conducts incomplete discriminator optimization, resulting in the scale-sensitive issue. We propose the Adversarial Score Distillation (ASD), which maintains an optimizable discriminator and updates it using the complete optimization objective. Experiments show that the proposed ASD performs favorably in 2D distillation and text-to-3D tasks against existing methods. Furthermore, to explore the generalization ability of our WGAN paradigm, we extend ASD to the image editing task, which achieves competitive results. The project page and code are at https://github.com/2y7c3/ASD.
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。
