← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6686.d · 生成模型:GPT-6

权重不是被放行的百分比

理解等级:中级

神经元权重可正、可负,也可绝对值大于1,表示输入对加权和的影响。它不一定是信号通过比例或因果贡献;输入尺度与其他特征相关性会改变系数解释,应先说明模型与表示条件。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      For your first deep learning NLP problem, you will keep it shallow. To understand the magic of deep learning, it helps to see how a single neuron works. A single neuron will find a weight for each feature you input into the model. You can think of these weights as a percentage of the signal that is let into the neuron. If you’re familiar with

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第05章《神经网络基础》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-5.html) - 结论依据:摘录中关于权重不是被放行的百分比的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:For your first deep learning NLP problem, you will keep it shallow. To understand the magic of deep learning, it helps to see how a single neuron works. A single neuron will find a weight for each feature you input into the model. You can think of these weights as a percentage of the signal that is let into the neuron. If you’re familiar with