Unverified50% confidenceFactExact time
FID（Fréchet Inception Distance）由Heusel等人于2017年提出，用预训练Inception网络提取特征并将分布建模为多元高斯，计算Fréchet距离，FID值越低表明分布越接近
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
Unverified可借助Fréchet Inception Distance（FID）或Maximum Mean Discrepancy（MMD）等分布距离度量来量化增强后训练分布与真实测试分布之间的偏移79% similarUnverifiedFID (Fréchet Inception Distance) evaluates generation quality by comparing the statistical distribution differences between real and generated data in the feature space of a deep neural network, with lower values indicating closer resemblance to real data.77% similarUnverified量化测量域偏移程度的常用方法包括最大均值差异（MMD）和Fréchet Inception Distance（FID）67% similarUnverifiedEthayarajh et al. revealed through empirical research in 2019 that word vectors in pre-trained models such as BERT and GPT are highly concentrated in a narrow cone-shaped region, exhibiting anisotropy.61% similarUnverifiedLi等人2018年发现神经网络的有效参数空间维度远低于名义参数量，训练轨迹被限制在低维子空间中60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/527102API
curl https://kongchang.com/api/v1/knowledge/claims/527102MCP
get_claim(id=527102)