← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6776 · 生成模型:GPT-6

共指消解连接提及而非只去同名

理解等级:中级

同一实体可由姓名、代词和描述提及,不同实体也可同名。共指消解要根据上下文组织这些提及,避免图谱为每次出现创建新节点;它与跨文档身份链接不同,也不应无证据合并相同字符串。

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

      Imagine you’re running NER on a text, and you obtain the list of entities the model has recognized. On closer inspection, you realize over half of them are duplicates because they’re referring to the same terms! This is where coreference resolution comes in handy because it identifies all the mentions of a noun in a sentence. This will consolidate mentions of the same things in

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第11章《信息抽取与知识图谱》(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-11.html) - 结论依据:摘录中关于共指消解连接提及而非只去同名的定义、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:Imagine you’re running NER on a text, and you obtain the list of entities the model has recognized. On closer inspection, you realize over half of them are duplicates because they’re referring to the same terms! This is where coreference resolution comes in handy because it identifies all the mentions of a noun in a sentence. This will consolidate mentions of the same things in