Unverified50% confidenceFactExact time
The statistical noise in SGD can help networks escape shallow local minima and saddle points, which is a systematic advantage in deep networks.
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
Unverified反向传播机制要求整个计算图必须是可微分的，且在极深网络中存在梯度消失/爆炸问题65% similarUnverifiedBatch normalization layers have highly noisy statistical estimates in small-batch scenarios, and combined with the region-level computation characteristics of Dice loss, this amplifies fluctuations in the predicted probability distribution.65% similarUnverified强行取极值样本会引入离群点，迫使SVM构建更复杂的决策边界，高阶poly核在这种数据上收敛更慢62% similarUnverified状态离散化存在权衡：分桶粒度过粗导致信息损失，过细则引发维度灾难，这也是DQN用神经网络替代Q表的核心动机62% similarUnverifiedVanishing gradient refers to the exponential decay of gradient values during backpropagation in deep neural networks, causing parameters near the input layer to receive almost no updates.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/556692API
curl https://kongchang.com/api/v1/knowledge/claims/556692MCP
get_claim(id=556692)