知识卡片
序列学习需说明顺序与重置
用序列模型表达学习器时,输入顺序和查询前的隐藏状态共同确定预测函数。集合任意指定顺序可构造序列,但不保证学习器对重排不变;比较独立测试查询还需说明是否从相同支持集状态重启,避免前一查询悄悄改变后一预测的条件。
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
- · 迁移学习与适配
原书出处 · 1 处
- · 查看原书依据
The goal of supervised learning is that after training on the training data, we should do well on any random test query. After processing the training sequence {x(1), y(1), …, x(n), y(n)}, an RNN will have arrived at some setting of its hidden state. Now if we feed the RNN a new query x(n+1), the mapping it applies, which produces an output y(n+1), can be
参考来源
- 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第37章《迁移学习与适配》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c037.xhtml)
- 结论依据:摘录中关于序列学习需说明顺序与重置的定义、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:The goal of supervised learning is that after training on the training data, we should do well on any random test query. After processing the training sequence {x(1), y(1), …, x(n), y(n)}, an RNN will have arrived at some setting of its hidden state. Now if we feed the RNN a new query x(n+1), the mapping it applies, which produces an output y(n+1), can be