提出iFID，让重建指标能准确预测扩散生成质量。
Making Reconstruction FID Predictive of Diffusion Generation FID
- 用潜在空间插值重构样本，计算与原图的FID差异。
- iFID与扩散生成FID相关性达0.85，显著优于传统指标。
- 适合评估VAE对扩散模型生成质量的影响，尤其关注泛化能力。
已知变分自编码器（VAE）的重建FID（rFID）与潜在扩散模型的生成FID（gFID）相关性差。本文提出插值FID（iFID），一种rFID的简单改进版本，能与gFID强相关。具体而言，对每个数据样本，检索其潜在空间最近邻，对两者潜码进行插值，解码后计算与原始样本的FID。我们从扩散泛化与幻觉现象角度解释iFID为何与gFID高度相关，且传统重建指标可能负相关。理论上，iFID衡量的是扩散采样集中在脊集附近的解码插值质量；实证上，iFID是首个在多种VAE上均与扩散生成FID呈现强相关性的指标，皮尔逊与斯皮尔曼相关系数约0.85。项目页面见https://tongdaxu.github.io/pages/ifid.html。
原文摘要 · Abstract (English)
It is well known that the reconstruction FID (rFID) of a VAE is poorly correlated with the generation FID (gFID) of a latent diffusion model. We propose interpolated FID (iFID), a simple variant of rFID that exhibits a strong correlation with gFID. Specifically, for each dataset element, we retrieve its nearest neighbor in latent space, interpolate between their latent representations, decode the interpolated latent, and compute the FID between the decoded samples and the original dataset. We provide an intuitive explanation for why iFID correlates well with gFID, and why reconstruction metrics can be negatively correlated with gFID, by connecting iFID to recent results on diffusion generalization and hallucination. Theoretically, we show that iFID evaluates decoded interpolations aligned with the ridge set around which diffusion samples concentrate, thereby measuring a quantity closely related to diffusion sample quality. Empirically, iFID is the first metric shown to strongly correlate with diffusion gFID across diverse VAEs, achieving Pearson and Spearman correlations of approximately $0.85$. The project page is available at https://tongdaxu.github.io/pages/ifid.html.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。