Building AI Agents with LLMs, RAG, and Knowledge Graphs — 书籍拆解
读到哪:未读。
readState不是 read/partial 的书不能当锚。
| 项 | |
|---|---|
| 作者 | Salvatore Raieli and Gabriele Iuculano |
| 版次 | 2025 版 |
| 格式 | epub ｜ 文本源 epub-builtin |
| 许可 | 自购/个人收藏 |
| 来源 | 主人个人藏书,2026-08 放入收件箱 |
| 清洗 | 删页眉页脚 0 行、页码 0 行、断词接回 1 处 |
我们重写的拆解(0 章)
（还没写。拆解是这本书对我们的真正产出——底下的元数据只是索引。）
为什么收它
记忆与知识这条线上的对照材料——尤其是「图谱 + 检索」这一格。
合法性
自购/个人收藏。来源:主人个人藏书,2026-08 放入收件箱。原始文件不入库,转码文本入库(私有库)。
出版方怎么说
（起草参考,不是我们的判断。真正的「覆盖什么/不覆盖什么」写进 frontmatter 的 claims / notCovered)
出版方简介(仅供起草参考,不是我们的判断):A practical guide to autonomous and modern AI agents
它覆盖什么、不覆盖什么
（还没读到能下判断的程度。claims / notCovered 空着就是空着,不猜。)
怎么引用它
(依据: book=building-ai-agents-rag-kg §Building AI Agents with LLMs, RAG, and Knowledge Graphs)
章节名对不上会被 lab:validate 拦下;页码锚(§p.123)同样可用。
结构(113 段,共 1039k 字符)
| 段 | 章节 | 页 | 规模 |
|---|---|---|---|
| 01 | Building AI Agents with LLMs, RAG, and Knowledge Graphs | — | 1.7k |
| 02 | Building AI Agents with LLMs, RAG, and Knowledge Graphs | — | 0.7k |
| 03 | Contributors | — | 1.3k |
| 04 | Building AI Agents with LLMs, RAG, and Knowledge Graphs | — | 1.5k |
| 05 | Table of Contents | — | 5.7k |
| 06 | Preface | — | 12.2k |
| 07 | Part 1: The AI Agent Engine: From Text to Large Language Models | — | 1.1k |
| 08 | 1 | — | 47.9k |
| 09 | 2 | — | 49.3k |
| 10 | 3 | — | 2.7k |
| 11 | The scaling law | — | 6.4k |
| 12 | Emergent properties | — | 2.6k |
| 13 | Context length | — | 1.7k |
| 14 | Mixture of experts | — | 50.9k |
| 15 | Part 2: AI Agents and Retrieval of Knowledge | — | 1.4k |
| 16 | 4 | — | 8.6k |
| 17 | The brain | — | 6.2k |
| 18 | The perception | — | 4.8k |
| 19 | Action | — | 20.3k |
| 20 | LangChain | — | 3.7k |
| 21 | Haystack | — | 2.9k |
| 22 | LlamaIndex | — | 1.9k |
| 23 | Semantic Kernel | — | 1.3k |
| 24 | AutoGen | — | 1.6k |
| 25 | Choosing an LLM agent framework | — | 13.8k |
| 26 | 5 | — | 17.5k |
| 27 | Chunking strategies | — | 8.6k |
| 28 | Embedding strategies | — | 9.5k |
| 29 | Embedding databases | — | 25.0k |
| 30 | 6 | — | 8.3k |
| 31 | Hierarchical indexing | — | 3.3k |
| 32 | Hypothetical questions and HyDE | — | 4.8k |
| 33 | Context enrichment | — | 1.8k |
| 34 | Query transformation | — | 1.9k |
| 35 | Keyword-based search and hybrid search | — | 2.5k |
| 36 | Query routing | — | 3.9k |
| 37 | Reranking | — | 9.4k |
| 38 | Response optimization | — | 10.3k |
| 39 | Training and training-free approaches | — | 14.4k |
| 40 | Data scalability, storage, and preprocessing | — | 7.8k |
| 41 | Parallel processing | — | 5.4k |
| 42 | Security and privacy | — | 23.1k |
| 43 | 7 | — | 6.4k |
| 44 | A formal definition of graphs and knowledge graphs | — | 5.6k |
| 45 | Taxonomies and ontologies | — | 4.5k |
| 46 | Knowledge creation | — | 8.7k |
| 47 | Creating a knowledge graph with an LLM | — | 5.6k |
| 48 | Knowledge assessment | — | 2.3k |
| 49 | Knowledge cleaning | — | 1.7k |
| 50 | Knowledge enrichment | — | 3.0k |
| 51 | Knowledge hosting and deployment | — | 7.3k |
| 52 | Graph-based indexing | — | 1.4k |
| 53 | Graph-guided retrieval | — | 8.8k |
| 54 | GraphRAG applications | — | 8.0k |
| 55 | Knowledge graph embeddings | — | 2.8k |
| 56 | Graph neural networks | — | 2.9k |
| 57 | LLMs reasoning on knowledge graphs | — | 15.8k |
| 58 | 8 | — | 13.4k |
| 59 | The multi-armed bandit problem | — | 10.0k |
| 60 | Markov decision processes | — | 13.5k |
| 61 | Model-free versus model-based approaches | — | 3.9k |
| 62 | On-policy versus off-policy methods | — | 3.4k |
| 63 | Exploring deep RL in detail | — | 25.4k |
| 64 | Challenges and future direction for deep RL | — | 33.2k |
| 65 | RL-enhanced LLMs | — | 1.9k |
| 66 | LLM-enhanced RL | — | 22.9k |
| 67 | Part 3: Creating Sophisticated AI to Solve Complex Scenarios | — | 1.2k |
| 68 | 9 | — | 18.7k |
| 69 | Toolformer | — | 2.6k |
| 70 | HuggingGPT | — | 18.0k |
| 71 | ChemCrow | — | 4.2k |
| 72 | SwiftDossier | — | 1.2k |
| 73 | ChemAgent | — | 3.8k |
| 74 | Multi-agent for law | — | 3.0k |
| 75 | Multi-agent for healthcare applications | — | 13.8k |
| 76 | Using HuggingGPT locally | — | 8.6k |
| 77 | Using HuggingGPT on the web | — | 29.7k |
| 78 | Software as a Service (SaaS) | — | 8.6k |
| 79 | Model as a Service (MaaS) | — | 12.2k |
| 80 | Data as a Service (DaaS) | — | 6.0k |
| 81 | Results as a Service (RaaS) | — | 6.9k |
| 82 | A comparison of the different paradigms | — | 16.6k |
| 83 | 10 | — | 4.5k |
| 84 | Starting with Streamlit | — | 6.6k |
| 85 | Caching the results | — | 8.4k |
| 86 | Adding the text elements | — | 2.5k |
| 87 | Inserting images in a Streamlit app | — | 2.3k |
| 88 | Creating a dynamic app | — | 27.7k |
| 89 | Model development | — | 6.8k |
| 90 | Model training | — | 7.7k |
| 91 | Model testing | — | 3.3k |
| 92 | Inference optimization | — | 17.6k |
| 93 | Handling errors in production | — | 2.3k |
| 94 | Security considerations for production | — | 18.4k |
| 95 | asyncio | — | 5.4k |
| 96 | Asynchronous programming and ML | — | 14.7k |
| 97 | Kubernetes | — | 1.5k |
| 98 | Docker with ML | — | 8.8k |
| 99 | 11 | — | 6.3k |
| 100 | Biomedical AI agents | — | 10.4k |
| 101 | Physical agents | — | 4.9k |
| 102 | LLM agents for gaming | — | 2.0k |
| 103 | Web agents | — | 1.4k |
| 104 | Challenges in human-agent communication | — | 8.7k |
| 105 | No clear superiority of multi-agents | — | 5.5k |
| 106 | Limits of reasoning | — | 10.0k |
| 107 | Creativity in LLM | — | 5.2k |
| 108 | Mechanistic interpretability | — | 12.7k |
| 109 | The road to artificial general intelligence | — | 5.3k |
| 110 | Ethical questions | — | 18.3k |
| 111 | Index | — | 22.3k |
| 112 | Why subscribe? | — | 3.9k |
| 113 | Contents | — | 6.3k |
我们自己的读书笔记(0 篇)
（还没有。读完某章后写进 docs/building-ai-agents-rag-kg/notes/,那才是这本书对我们的产出。）
本页由 node scripts/book-build.mjs 生成:表格来自转码结果,散文来自书卡正文。不要手改本页。