AI認知與應用通識教程

第一篇 AI認知與大模型基礎
第1章 人工智能通識：從“神秘”到“日常” …………………………………………002
1.1 你已經在用AI了——生活中無處不在的人工智能 ···········································002
1.2 到底什麼是人工智能？——理解AI的三個角度 ··············································004
1.3 AI的“能力範圍”——專用與通用 ·······························································006
1.4 AI技術體系——主要分支速覽 ······································································007
第2章 大模型基礎：從“識字”到“懂意” …………………………………………014
2.1 大模型如何“讀”你的話——從文字到數字 ···················································014
2.2 大模型如何“抓重點”——註意力機制與上下文 ··················································019
2.3 大模型的“記憶”——它記得什麼、忘了什麼 ················································020
2.4 大模型是怎麼“煉成”的——訓練的三個階段 ················································023
2.5 主流大模型一覽——它們有什麼不同 ·····························································025
第3章 大模型能力：從“聽懂”到“創造” …………………………………………029
3.1 模型的“五官”：多模態 ·············································································029
3.2 核心解剖：大模型的五大“超能力” ·····························································032
3.3 定制你的專屬模型：微調與RAG ··································································035
3.4 能力邊界與安全考量：不要神化AI ·······························································038
第4章 大模型安全：從“幻覺”到“防護” …………………………………………040
4.1 幻覺：當模型“自信地胡說” ·······································································040
4.2 攻擊方式：壞人如何“忽悠”模型 ································································045
4.3 防護手段：給模型穿上“防彈衣” ································································049
4.4 AI倫理簡論：技術背後的價值選擇 ·······························································054
第二篇 提示詞設計與高效表達
第5 章 提示詞工程基礎：從“問對問題”開始 ……………………………………… 059
5.1 為什麼提示詞決定答案質量 ·········································································.059
5.2 基礎技巧：讓AI“聽懂你” ········································································.064
5.3 Markdown 基礎：提示詞的“顏值即正義” ··················································.068
5.4 實戰：從“糟糕提示詞”到“優秀提示詞” ··················································.073
5.5 評估與疊代：如何知道提示詞“好”了 ·························································.078
第6 章 結構化提示詞：用結構化語言為AI“編程” ………………………………… 082
6.1 結構化提示詞的優勢與用處 ·········································································.082
6.2 JSON 基礎：AI 時代的“通用語” ·······························································.085
6.3 XML 基礎：另一種結構化選擇 ····································································.088
6.4 實戰：結構化提示詞應用場景 ······································································.091
6.5 解析與驗證AI 返回的結構化結果·································································.094
6.6 結構化提示詞的局限與未來 ·········································································.096
第7 章 提示詞思維方法論：從“技巧應用”到“思維升級” ……………………… 100
7.1 為什麼需要思維方法論 ················································································.100
7.2 SCQA 框架：講好一個“故事” ···································································.102
7.3 5W1H 框架：覆蓋所有維度 ·········································································.106
7.4 實戰：設計高階提示詞模板 ·········································································.109
7.5 綜合應用：構建你的提示詞設計體系 ······························································111
7.6 未來：提示詞即思維訓練 ············································································.114
第三篇 知識管理與RAG 應用
第8 章 RAG 技術：讓大模型“有據可查” …………………………………………… 117
8.1 RAG 是什麼 ·······························································································.117
8.2 RAG 背後的關鍵技術 ·················································································.122
第9 章 個人知識庫構建：打造你的“第二大腦” …………………………………… 127
9.1 為什麼需要構建個人知識庫 ·········································································.127
9.2 知識庫構建的四階段 ····················································································130
9.3 知識庫×AI：從“存儲”到“智能” ····························································136
9.4 長期維護：讓知識庫“越長越聰明” ·····························································138
第四篇 Agent設計與OpenClaw基礎
第10章 認識Agent：能力、價值與局限 ………………………………………………141
10.1 前Agent時代：從搜索引擎到ChatGPT ······················································141
10.2 什麼是Agent？——不只是聊天機器人 ·······················································143
10.3 多Agent協作：從“單打獨鬥”到“分工協作” ··········································145
10.4 Agent的局限與風險 ··················································································148
第11章 Agent如何想與記 ………………………………………………………………153
11.1 Agent的思維鏈 ·························································································153
11.2 ReAct模式：邊想邊做邊觀察 ·····································································154
11.3 規劃：把大任務拆成小步驟 ········································································157
11.4 記憶：Agent的“經驗庫” ········································································158
第12章 Agent如何調用工具 ……………………………………………………………162
12.1 Agent的“工具箱” ··················································································162
12.2 工具調用中的常見概念解釋 ········································································165
12.3 工具調用流程：以“智能生活助手”為例 ·····················································167
第13章 Agent項目開發：從需求到測試的完整旅程 …………………………………174
13.1 項目啟動：從“一個好想法”開始 ······························································174
13.2 設計階段：畫出智能體的“藍圖” ······························································177
13.3 實現階段：讓Agent“活”起來 ··································································179
13.4 測試階段：確保Agent“不闖禍” ·······························································184
13.5 項目總結與展示 ·························································································188
第14章 OpenClaw：Agent的“協作與創造”新範式………………………………191
14.1 Agent協作體系的運行框架與OpenClaw技術基礎 ········································191
14.2 OpenClaw部署與應用基礎 ·········································································194
14.3 OpenClaw核心功能拆解 ·············································································195
14.4 OpenClaw 的企業級落地優化 ·····································································.200
第15 章 總結與展望：AI 的過去、現在與未來 ………………………………………… 204
15.1 全書知識體系回顧 ····················································································.204
15.2 核心概念梳理：一張表看懂AI ···································································.205
15.3 AI 未來展望：技術趨勢與行業變革 ····························································.207
15.4 結語：AI 時代，你的位置在哪裏 ································································.208
附錄A 使用OpenClaw 制作PPT 實例 ……………………………………………… 210
A.1 PPT 生成相關技能安裝及內置原生工具介紹 ·············································210
A.2 PPT 生成實例 ···························································································.212
A.3 使用OpenClaw 生成PPT 的內在實現機制分析 ·············································.214
小結 ·················································································································.215
附錄 B O penClaw官方資源導航（截至2026年7月） …………………………………216
B.1 官方網站 ···································································································.216
B.2 官方開發資源 ····························································································.216
B.3 官方技能生態 ····························································································.217
B.4 官方推薦開發工具······················································································.217
B.5 學習建議 ···································································································.217
B.6 使用建議 ···································································································.218
附錄C 非技術人員的AI 學習指南 ……………………………………………………… 219
附錄D 國內AI 工具訪問地址匯總 ……………………………………………………… 222
參考文獻……………………………………………………………………………………… 224