Unverified50% confidenceFactExact time
Random weight initialization breaks the symmetry between neurons, so that each neuron starts from a different point and thus explores different feature combinations.
1
Sources
50%
Confidence
Long-term
Relevance
10/23/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFor each generator, the key neurons consistently selected account for less than 1% of the total, and they are highly consistent across different cross-validation folds and random seeds.77% similarUnverified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76% similarUnverifiedAll-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.73% similarUnverifiedWhen a layer of Inception V1 was cut from 480 neurons to 10 and retrained, the network did not exhibit superposition; instead, neuron functions became coarser and more generalized.72% similarUnverified深度学习中的随机种子控制权重初始化、数据加载批次打乱顺序、Dropout神经元随机失活以及优化器中的随机梯度估计等多个随机性来源70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1087983API
curl https://kongchang.com/api/v1/knowledge/claims/1087983MCP
get_claim(id=1087983)