arXiv：cs.LG· Yunghee Lee, Jaeyeon Kim·· 4 小时前AI 评分44
什么是 Fréchet 距离？一种方向性分解方法
What Does Fr\'echet Distance Measure? A Directional Decomposition
AI 导读
研究提出方向性 Fréchet 距离，将最优传输位移在给定方向上的平方投影期望作为度量，让 FID/FVD 等指标可解释化。在图像、视频和蛋白质案例中，少量可解释方向即可解释大部分距离，并借此用 CLIP 嵌入的语义概念解释 COCO 上扩散采样步数增加导致 ImageReward 提升但 FID 变差的现象，同时量化 FVD 对逐帧外观的偏好。代码已开源。
正文
Abstract:The Fréchet distance is a de facto standard for evaluating generative models across domains, appearing as FID for images and FVD for videos. It summarizes the discrepancy between generated and reference distributions in a single scalar, with lower values typically interpreted as better generation quality. However, this scalar view can obscure what drives the comparison. For example, in COCO dataset, increasing the number of diffusion sampling steps improves ImageReward scores yet worsens (increases) FID. Motivated by this mismatch, we seek to make the Fréchet distance more interpretable by uncovering where the discrepancy lies. To this end, we introduce directional Fréchet distance, the expected squared projection of the optimal transport displacement onto a given direction. Across our image, video, and protein case studies, we find that a small number of interpretable directions account for much of the distance. We use these directions to explain the FID increase in terms of semantic concepts represented by CLIP embeddings, quantify FVD's bias toward per-frame appearance, and revisit the interpretation of Protein FID. We open-source our codebase at this https URL.
| Comments: | 20 pages, 4 figures, 7 tables |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.05518 [cs.AI] |
| (or arXiv:2610.05518v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.05518 arXiv-issued DOI via DataCite |
Submission history
From: Yunghee Lee [view email]
[v1]
Sun, 4 Oct 2026 20:35:04 UTC (252 KB)
来源：arXiv：cs.LG · arxiv.org