Unverified50% confidenceCautionExact time
All-zero initialization causes a symmetry problem: all neurons in the same layer compute identical gradients and make identical updates, making it impossible to break symmetry.
1
Sources
50%
Confidence
Long-term
Relevance
8/30/2026
First Seen
Sources
Related Entities
Related Claims
Unverified叠加现象指神经网络倾向于用远少于特征数量的神经元同时编码大量语义特征，与传统一个神经元一个特征的直觉相悖61% similarUnverifiedThe residual structure makes F=0 the default option—as long as the F(x) weights are close to zero, the residual block degenerates into an identity mapping, thereby resolving the optimization difficulty of SGD in learning identity mappings.60% similarVerifiedThe vanishing gradient problem originates from saturating activation functions such as Sigmoid or Tanh, whose derivatives approach zero for large or small input values, causing gradients to decay layer by layer.60% similarUnverifiedsigmoid激活函数的导数最大值仅为0.25，多层连乘后导致梯度迅速消失59% similarUnverified仅无约束地最大化SWD会导致神经网络扭曲数据结构得到退化解，因此需引入几何约束在分离性与保真度之间权衡59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/826866API
curl https://kongchang.com/api/v1/knowledge/claims/826866MCP
get_claim(id=826866)