知识卡片
词表覆盖限制静态类比答案
若目标答案没有对应词项,纯静态向量近邻不会凭空产生该答案。训练语料很大仍不保证每个旧词、专名或短语进入最终词表,低频过滤和预处理都影响覆盖;使用前需实际检查键及词表版本。
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
- · 用词嵌入推理
原书出处 · 1 处
- · 查看原书依据
Word embeddings can be used to answer even these vague questions and analogy problems, and they can help you remember any word or name on the tip of your tongue, as long as the vector for the answer exists in your vocabulary. (For Google’s pretrained Word2Vec model, your word is almost certainly within the 100-billion word news feed that Google trained it on, unless your
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:摘录中关于词表覆盖限制静态类比答案的定义、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:Word embeddings can be used to answer even these vague questions and analogy problems, and they can help you remember any word or name on the tip of your tongue, as long as the vector for the answer exists in your vocabulary. (For Google’s pretrained Word2Vec model, your word is almost certainly within the 100-billion word news feed that Google trained it on, unless your