😇

Building AI Search: Multi-Modal RAG, RAFT, & GraphRAG

Course Introduction and Logistics

Module 2: Building AI Search: Key Components of Advanced RAG

Introduction to Advanced RAG Rerankers: Supercharging your retrieval results Hybrid search: Combining the strengths of dense and sparse vectors Restaurant Reviews: using Hybrid search with Deep Lake Advanced chunking: Moving beyond arbitrary token chunking Fine-tuning: Adapting embeddings to your specific domain Multimodal RAG: Retrieving Images & More Restaurant Insights: Multimodal RAG with Deep Lake Multi-modal AI Search Across Restaurants with ColPali Notable Techniques: ColBERT & Contextual Retrieval Using PaperQA2 for Scientific Discovery

Module 3: Graphs and Retrieval Augmented Generation

Introduction to GraphRAG Graph RAG and Vector Search for AI Recipe Discovery Distill-SynthKG by Intel Labs and Salesforce Research: Distilling Knowledge Graph Synthesis Workflow for Improved Coverage and Efficiency Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering by LinkedIn Corporation

Module 4: RAFT (Retrieval Augmented Fine-Tuning)

Module Introduction: The Best of Both Worlds with RAFT

RAFT Video by LlamaIndex: https://www.youtube.com/watch?v=sqPckknlgDc

Practical Project Overview: Using TorchTune and Deep Lake for RAFT

Module 5: Evaluating RAG Systems

Introduction to RAG evaluation Evaluation dataset: Setting up the target correctly Evaluation metrics: From ROUGE to LLMs RAG evaluation tools: Overview of the open-source eval landscape Optimizing RAG on ai-arxiv Dataset Conclusion