知识卡片

普通读书笔记卡 · 6686.a · 生成模型:GPT-6

单层感知器只能线性划分输入

理解等级:中级

加权和后取硬阈值形成一个线性决策边界,不能在原二维输入上正确分开XOR。增加合适中间非线性表示可扩大表达能力,但不是任何单个神经元都能逼近任意函数,也不保证给足时间就可解决全部学习问题。

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

      AI researchers hoped to replace the rigid math of logistic regressions, linear regressions, and polynomial feature extraction with the more fuzzy and generalized logic of neural networks—tiny brains. Rosenblatt’s artificial neurons even worked for trigonometric functions and other highly nonlinear functions. Each neuron solved one part of the problem and could be combined with other neurons to learn more and more complex functions (though not

    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:摘录中关于单层感知器只能线性划分输入的定义、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:AI researchers hoped to replace the rigid math of logistic regressions, linear regressions, and polynomial feature extraction with the more fuzzy and generalized logic of neural networks—tiny brains. Rosenblatt’s artificial neurons even worked for trigonometric functions and other highly nonlinear functions. Each neuron solved one part of the problem and could be combined with other neurons to learn more and more complex functions (though not