知识卡片
循环状态与注意力结构要分开
RNN通过状态沿序列传递信息,LSTM与GRU是门控循环结构。Transformer以注意力为核心,原始结构明确避免循环,不能简单把它列为RNN的变种;本章仅概览,不补造未展开训练细节。
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
- · 深度学习模型
原书出处 · 1 处
- · 查看原书依据
In addition to MLPs and CNNs, a variety of other neural network architectures have been developed for different types of data. One of the most prevalent of these architectures is the RNN. RNNs are designed to operate most effectively on time-series data (e.g., a data stream read at sequential time stamps from a sensor). Unlike traditional feedforward models such as MLPs and CNNs, recurrent layers
参考来源
- 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第08章《深度学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/21_chapter08.xhtml)
- 结论依据:摘录中关于循环状态与注意力结构要分开的定义、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:In addition to MLPs and CNNs, a variety of other neural network architectures have been developed for different types of data. One of the most prevalent of these architectures is the RNN. RNNs are designed to operate most effectively on time-series data (e.g., a data stream read at sequential time stamps from a sensor). Unlike traditional feedforward models such as MLPs and CNNs, recurrent layers
- 外部纠错依据(2026-11-06核验):https://arxiv.org/abs/1706.03762;仅用于限定原书错误,不替代本卡原书出处。