Unverified50% confidenceFactExact time
反向传播的四个基本方程围绕误差δˡⱼ=∂C/∂zˡⱼ推导：输出层误差δᴸ=∇ₐC⊙σ′(zᴸ)、误差反向传播、偏置梯度∂C/∂bˡ=δˡ、权重梯度∂C/∂wˡ=δˡ(aˡ⁻¹)ᵀ
1
Sources
50%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
梯度下降与反向传播详解：神经网络学习机制从零拆解
Related Claims
UnverifiedThe residual block formula is y = x + F(x); during backpropagation, the gradient passes through the "1" in (1 + dF/dx), providing a shortcut for gradients to flow directly to earlier layers.59% similarUnverifiedDiffusionGemma adopts the discrete diffusion path.57% similarUnverified墨卡托投影的坐标变换公式为 x = R·λ，y = R·ln[tan(π/4 + φ/2)]56% similarPartially VerifiedThe mathematical form of KL divergence is D_KL(p||q) = ∑ p(x_i) log(p(x_i)/q(x_i)).53% similarUnverified当球面信号被旋转时，其第l阶球谐系数向量会被一个(2l+1)×(2l+1)的Wigner-D矩阵所变换，不同阶之间互不耦合，旋转在频域中呈现块对角结构53% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/558145API
curl https://kongchang.com/api/v1/knowledge/claims/558145MCP
get_claim(id=558145)