Unverified50% confidenceFactExact time
重要性采样的数学基础是Radon-Nikodym定理与测度变换，通过似然比w=p(x)/q(x)对样本重加权实现无偏估计
1
Sources
50%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
引导式生成模型：精准估算极端事件概率的新方法
Related Claims
Unverified重要性采样的数学核心是测度变换，通过引入提议分布q并用似然比w=p(x)/q(x)对样本重加权，无偏地估算原始概率82% similarUnverifiedThe posterior distribution is obtained via Bayes' theorem: p(θ|x) ∝ p(x|θ)·p(θ), combining prior knowledge with the data likelihood.66% similarUnverified重参数化技巧将采样表示为 z = μ + σ × ε（ε从标准正态分布采样），使随机性外置于ε，从而让μ和σ的梯度可正常反向传播实现端到端训练66% similarUnverifiedThe differentiability of Monte Carlo PDE solvers stems from the fact that their estimators are themselves smooth functions of the parameters, allowing unbiased gradient estimates to be obtained directly via reparameterization tricks or likelihood ratio methods, without needing to backtrack through the intermediate states of a large-scale linear system.62% similarUnverifiedPyTorch判断连续性的公式是：对于n维张量，如果stride[i] == stride[i+1] * size[i+1]对所有i成立，则该张量连续60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/510571API
curl https://kongchang.com/api/v1/knowledge/claims/510571MCP
get_claim(id=510571)