Unverified50% confidenceFactExact time
当需要表征的特征数量远超可用神经元数量时，网络会利用高维空间中近似正交向量数量远超维度数量的特性（Johnson-Lindenstrauss引理）将多个特征叠加压缩编码到同一组神经元上
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
Unverified深度神经网络由输入层、若干隐藏层、输出层构成，每个神经元对前一层输入做加权求和加偏置后通过激活函数做非线性变换78% similarUnverified大模型本质上是一个由深度神经网络实现的极其复杂的函数72% similarVerifiedDeep learning uses multi-layer neural networks whose structure is inspired by the way neurons in the human brain are connected.72% similarUnverifiedTelgarsky等人的工作证明，深度网络在表示某些函数时比等宽度的浅层网络具有指数级的参数效率优势69% similarUnverifiedLarge language models are essentially deep neural networks.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/313862API
curl https://kongchang.com/api/v1/knowledge/claims/313862MCP
get_claim(id=313862)