Скачать или смотреть LLM Chronicles #:6.2 RAG (Self-Query/Parent Document/HyDE) - Part 2

This episode explores some issues with RAG, covering solutions and common strategies to mitigate them and improve performance such as self-query, parent document retrieval and HyDE (Hypothetical Document Embedding).

Series website: https://llm-chronicles.com/

🖹 Download the mindmap for this episode here:
Retrieval Augmented Generation - Full: https://llm-chronicles.com/pdfs/llm-c...

🖹 Lab Notebooks:
https://colab.research.google.com/dri...
https://colab.research.google.com/dri...
https://colab.research.google.com/dri...

🕤 Timestamps:
00:00 - Introduction
00:33 - Training Sentence Similarity Embedding Models
02:26 - Self-Query Retriever
04:43 - Text Segmenetation
05:20 - Size of Chunks
06:01 - Parent Document Retriever
07:04 - Hypothetical Document Embeddings (HyDE)
07:48 - Lab 1 - Self-Query Retriever in Langchain
14:32 - Lab 2 - Parent Document Retriever in Langchain
18:56 - Lab 3 - HyDE in Langchain

📚 References & Acknowledgements:

https://arxiv.org/abs/2312.10997, "Retrieval-Augmented Generation for Large Language Models: A Survey"
https://arxiv.org/abs/2212.10496, "Precise Zero-Shot Dense Retrieval without Relevance Labels"