Unverified50% confidenceFactExact time

对比学习会产生对齐性（alignment）与均匀性（uniformity）两种相互竞争的几何属性

1

Sources

50%

Confidence

Long-term

Relevance

7/9/2026

First Seen

Sources

Block-Sparse Featurizers: Opening the Black Box of Vision Model Neural Geometry

hackernewshackernews7/7/2026

Related Claims

UnverifiedCLIP 用对比学习（Contrastive Learning）替代了此前的监督分类范式，通过最大化匹配图文对相似度、最小化不匹配对相似度进行自监督学习65% similar Unverified滚动截图拼接的核心难点在于重叠区域检测，常见方案通过归一化互相关系数或差值平方和进行相似度比对来找到最佳拼接点64% similar Unverified余弦相似度衡量的是两个向量在高维球面上的角度差异，依赖全局方向信息，忽略嵌入向量各维度之间的局部关系结构61% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/322194

API

curl https://kongchang.com/api/v1/knowledge/claims/322194

MCP

get_claim(id=322194)