Avrilyn_n
@AvrilynLi
AI 🌍🇨🇳🇨🇦🇺🇸
加入 August 2022
502 正在關注 956 粉絲
8 种 RAG 架构，AI 工程师必懂（附适用场景）
1）Naive RAG
基于 Query 与文档向量的相似度进行检索。
适合事实查询、知识问答等简单场景，语义匹配即可满足需求。
2）Multimodal RAG（多模态 RAG）
同时检索文本、图片、音频等多种数据。
适用于跨模态任务，例如根据文字问题结合图片和文本进行回答。
3）HyDE（Hypothetical Document Embeddings）
当 Query 与目标文档语义差距较大时效果更好。
先让模型根据 Query 生成一篇“假想答案”，再对其进行向量化检索。
利用假想文档的语义表示找到更相关的真实文档。
4）Corrective RAG（校正型 RAG）
对检索结果进行二次验证，例如结合 Web Search 等可信来源。
在送入 LLM 前过滤、修正过时或错误信息，提高回答准确性。
5）Graph RAG
将检索结果组织成知识图谱，显式表示实体及其关系。
为 LLM 提供结构化上下文，增强推理能力，尤其适合复杂知识关联场景。
6）Hybrid RAG（混合 RAG）
将向量检索与图检索结合到同一流程中。
同时利用非结构化文本和结构化关系数据，获得更完整、更准确的答案。
7）Adaptive RAG（自适应 RAG）
根据问题复杂度动态选择检索策略。
简单问题直接检索，复杂问题自动拆分为多个子问题进行多轮检索与推理。
8）Agentic RAG（Agent RAG）
引入 AI Agent，通过规划、推理（如 ReAct、CoT）和长期记忆协调检索流程。
可以调用工具、API、多数据源，甚至组合多种 RAG 方法完成复杂任务。
大多数 RAG 架构的核心区别，都发生在检索阶段（Retrieval Time）：决定检索什么、检索多少、如何组合信息。
但它们都建立在同一个前提上——索引（Indexing）已经做好了。
如果索引阶段切分的质量很差，无论采用哪种 RAG 架构，最终效果都会受到限制。优化索引质量，本身就是一个独立且非常重要的问题。
Akshay写了一种新的 Indexing 方法，声称可以做到：
将语料规模减少约 40 倍。
单次查询 Token 消耗降低约 3 倍。
向量检索相关性提升约 2.3 倍。
而且，它无需修改 Retrieval 算法、Reranker 或 Embedding 模型，只优化了 Indexing 本身。
顯示更多
8 RAG architectures for AI Engineers:
(explained with usage)
1) Naive RAG
- Retrieves documents purely based on vector similarity between the query embedding and stored embeddings.
- Works best for simple, fact-based queries where direct semantic matching suffices.
2) Multimodal RAG
- Handles multiple data types (text, images, audio, etc.) by embedding and retrieving across modalities.
- Ideal for cross-modal retrieval tasks like answering a text query with both text and image context.
3) HyDE (Hypothetical Document Embeddings)
- Queries are not semantically similar to documents.
- This technique generates a hypothetical answer document from the query before retrieval.
- Uses this generated document’s embedding to find more relevant real documents.
4) Corrective RAG
- Validates retrieved results by comparing them against trusted sources (e.g., web search).
- Ensures up-to-date and accurate information, filtering or correcting retrieved content before passing to the LLM.
5) Graph RAG
- Converts retrieved content into a knowledge graph to capture relationships and entities.
- Enhances reasoning by providing structured context alongside raw text to the LLM.
6) Hybrid RAG
- Combines dense vector retrieval with graph-based retrieval in a single pipeline.
- Useful when the task requires both unstructured text and structured relational data for richer answers.
7) Adaptive RAG
- Dynamically decides if a query requires a simple direct retrieval or a multi-step reasoning chain.
- Breaks complex queries into smaller sub-queries for better coverage and accuracy.
8) Agentic RAG
- Uses AI agents with planning, reasoning (ReAct, CoT), and memory to orchestrate retrieval from multiple sources.
- Best suited for complex workflows that require tool use, external APIs, or combining multiple RAG techniques.
Most architectures here involve some form of retrieval-time decision. But they all run on top of whatever was already indexed.
If that indexing step outputs messy chunks, every architecture inherits them. Improving it is a separate problem from the 8 above.
I wrote about a better unit for the indexing step. The technique:
- cuts corpus size by 40x.
- reduces tokens per query by 3x.
- improves vector search relevance by 2.3x.
And it doesn't alter the retrieval algorithm, the reranker, or the embedding model.
Read it in the article quoted below.
顯示更多
