Unverified50% confidenceFactExact time
The weight gradient formula is ∂L/∂W = δ · a_prev^T.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Building a Neural Network from Scratch with NumPy: A Hands-On Deep Dive into Backpropagation
redditr/deeplearning10/19/2026
Related Entities
Related Claims
UnverifiedThe typical form of Zeller's formula is h = (q + ⌊13(m+1)/5⌋ + K + ⌊K/4⌋ + ⌊J/4⌋ - 2J) mod 7.71% similarUnverifiedThe closed-form solution for linear regression coefficients is β1 = Σ(xi-x̄)(yi-ȳ)/Σ(xi-x̄)², β0 = ȳ - β1·x̄69% similarUnverified含时薛定谔方程的形式为 iℏ∂ψ/∂t = Ĥψ，其中ℏ是约化普朗克常数，Ĥ是哈密顿算符67% similarUnverified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)ᵀ66% similarUnverifiedThe proximal operator prox_{ηh}(v) = argmin_x { h(x) + (1/(2η))||x-v||^2 } is equivalent to performing one step of implicit gradient descent on h at the point v, and does not require h to be differentiable.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1069364API
curl https://kongchang.com/api/v1/knowledge/claims/1069364MCP
get_claim(id=1069364)