Building LLM Applications: Advanced RAG (Part 10)
Learn Large Language Models ( LLM ) through the lens of a Retrieval Augmented Generation ( RAG ) Application.
Posts in this Series
- Introduction
- Data Preparation
- Sentence Transformers
- Vector Database
- Search & Retrieval
- LLM
- Open-Source RAG
- Evaluation
- Serving LLMs
- Advanced RAG ( This Post )
Table Of Contents
· 1. Naive RAG
· 1.1. Issues with Naive RAG
∘ 1.1.1. Indexing
∘ 1.1.2. Retrieval
∘ 1.1.3. Generation
· 2. Move Towards Advanced RAG
∘ 2.1. Pre-Retrieval Optimization
∘ 2.2. Retrieval optimization
∘ 2.3. Post-retrieval optimization
· 3. Advanced RAG Techniques
· 4. PRE-RETRIEVAL TECHNIQUES
· 4.1. PDF Parsing
∘ 4.1.1. The Challenges of Parsing PDF
∘ 4.1.2. How to parse PDF documents
∘ 4.1.3. Rule-based methods
∘ 4.1.4. Methods based on deep learning models.
· 4.2. Context Enrichment
∘ 4.2.1. Sentence Window Retrieval
∘ 4.2.2. Auto-merging Retriever (aka Parent Document Retriever)
· 4.3. Query Rewriting
∘ 4.3.1. Hypothetical Document Embeddings (HyDE)
∘ 4.3.2. Rewrite-Retrieve-Read
∘ 4.3.3. Step-Back Prompting
∘ 4.3.4. Query2doc
∘ 4.3.5. ITER-RETGEN
· 4.4. Semantic Chunking
∘ 4.4.1. Embedding-based methods
∘ 4.4.2. Model-based methods
∘ 4.4.2.1. Naive BERT
∘ 4.4.2.2. Cross Segment…