Auto-sync: 2026-04-17 13:17
This commit is contained in:
30
wiki/concepts/Knowledge-Extraction.md
Normal file
30
wiki/concepts/Knowledge-Extraction.md
Normal file
@@ -0,0 +1,30 @@
|
||||
---
|
||||
title: "Knowledge Extraction"
|
||||
type: concept
|
||||
tags: [ai, knowledge-management, automation]
|
||||
sources: [self-healing-home-server-infrastructure-management]
|
||||
last_updated: 2026-04-17
|
||||
---
|
||||
|
||||
## Summary
|
||||
Knowledge Extraction(知识提取)是指 AI Agent 从非结构化或半结构化数据源(如笔记、对话记录、邮件)中提取有价值信息并转化为结构化、可查询知识库的过程。
|
||||
|
||||
## Definition
|
||||
将分散的、非结构化的信息通过 AI 处理转化为结构化、可复用知识的技术。
|
||||
|
||||
## Key Characteristics
|
||||
- 从多种来源提取信息(Obsidian 笔记、ChatGPT 对话导出、邮件)
|
||||
- 转化为原子事实(atomic facts)
|
||||
- 构建可搜索的知识库
|
||||
- 随着时间推移价值递增
|
||||
|
||||
## Benefits
|
||||
- 将个人知识资产结构化
|
||||
- 创建可搜索的知识库
|
||||
- 随时间累积价值(一个用户从 ChatGPT 历史中提取了 49,079 个原子事实)
|
||||
- 支持 AI Agent 的上下文记忆
|
||||
|
||||
## Connections
|
||||
- [[Obsidian]] ← stores ← [[Knowledge Extraction]]:Obsidian 笔记是知识提取的来源之一
|
||||
- [[Agentic AI]] ← performs ← [[Knowledge Extraction]]:AI Agent 执行知识提取
|
||||
- [[Knowledge Network]] ← built_by ← [[Knowledge Extraction]]:知识提取构建知识网络
|
||||
Reference in New Issue
Block a user