← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9170.a · 生成模型:GPT-6

前向传播区分线性量和输出

理解等级:中级

每个节点先由前层输出加权求和并加偏置,得到预激活a,再经g(a)产生输出。层号、源节点与目标节点索引必须一致;不同文献把activation用于不同量,实现应按实际公式而非术语猜测。

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

      where g() is a nonlinear activation function such as (but not limited to) the ones described in section 10.2 or the Heaviside step function. Here, is known as the activation—that is, the weighted sum computed by node i in layer k:

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第10章《人工神经网络》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_10.xhtml) - 结论依据:摘录中关于前向传播区分线性量和输出的定义、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:where g() is a nonlinear activation function such as (but not limited to) the ones described in section 10.2 or the Heaviside step function. Here, is known as the activation—that is, the weighted sum computed by node i in layer k: