Unverified50% confidenceTradeoffExact time
A single hidden layer is not inferior to deep networks in expressive power, but deep structures can express the same family of functions with fewer parameters, thereby affecting trainability.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedLow-level features such as edges and textures learned by the shallow layers of deep learning models exhibit strong domain-agnostic properties, while high-level semantic features encoded by deep layers are more domain-specific.77% similarUnverified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73% similarUnverifiedThe 'depth' in deep learning refers to a network containing multiple hidden layers.72% similarUnverifiedSPLADE and DeepImpact are representatives of learned sparse retrieval, producing sparse vectors whose per-term weights are learned by a neural network and enabling term expansion.70% similarUnverified残差连接与层归一化解决了深层网络训练不稳定的问题，使模型可堆叠到数十层69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1077981API
curl https://kongchang.com/api/v1/knowledge/claims/1077981MCP
get_claim(id=1077981)