知识卡片
记忆机制并不只有循环网络
RNN通过隐藏状态提供记忆,固定历史窗口、显式状态机等也能利用过去信息。选择应看所需历史长度、可观性、训练数据与在线延迟;需要状态不等于必须使用本章的一种网络架构。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 人工神经网络
原书出处 · 1 处
- · 查看原书依据
• When a system is not purely reactive but requires state (described in chapter 11), recurrent neural networks are needed to implement a notion of memory. Exercises 1. Implement the simple perceptron training algorithm and use it to find a separating hyperplane for simple data. 2. Find out how to implement the auto differentiation (or auto gradient) function in your favorite numerical package (e.g., NumPy
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:摘录中关于记忆机制并不只有循环网络的定义、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:• When a system is not purely reactive but requires state (described in chapter 11), recurrent neural networks are needed to implement a notion of memory. Exercises 1. Implement the simple perceptron training algorithm and use it to find a separating hyperplane for simple data. 2. Find out how to implement the auto differentiation (or auto gradient) function in your favorite numerical package (e.g., NumPy