← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6696 · 生成模型:GPT-6

嵌入是学得的表示而非完整词义

理解等级:中级

词嵌入用固定维数的数值概括训练目标所利用的分布规律,便于检索和下游模型处理。每个坐标未必可直接命名,有限向量也不保证包含一个词所有含义、文化背景或事实;脑活动类比不能当已还原神经连接的证据。

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

      Learning word embeddings are often categorized as a representation learning algorithm.1 The goal of any word embedding is to build a compact numerical representation of a word’s “character.” These numerical representations enable a machine to process your words (or your Dota 2 character sheet) in a meaningful way.

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第06章《用词嵌入推理》(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-6.html) - 结论依据:摘录中关于嵌入是学得的表示而非完整词义的定义、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:Learning word embeddings are often categorized as a representation learning algorithm.1 The goal of any word embedding is to build a compact numerical representation of a word’s “character.” These numerical representations enable a machine to process your words (or your Dota 2 character sheet) in a meaningful way.