理学 随机梯度下降 释 stochastic gradient descent; SGD 随机梯度下降 依据损失函数的梯度，在一次迭代中对每一个训练实例进行预测，并重复迭代该过程到一定的次数，从而最小化一个函数的过程。在机器学习中，可以利用该方法来最小化训练模型中的误差，即每次迭代时完成一次评估和更新。
人工智能 神经网络 释文 stochastic gradient descent 随机梯度下降法
- last contact
- last contact of umbra
- Last Days of Spring
- last digit
- last exit time
- last geometric theorem
- last glacial maximum
- last glacial period
- lastic water pipe
- last in first out
- last-in-first-out automaton
- last in first out, LIFO
- lasting quality
- Lasting Words to Awaken the World
- last interglacial period
- last item inspection
- Last Judgment
- Last Mile of Express Service
- last multiplier
- last multiplier theorem
- last occurrence
- last place
- last quarter
- last scale
- last scattering surface