vector-store

• “vector-store”
  • “rag”
  • “knowledge-graph”
  • “neo4j”
  • “graph-rag”
  • “semantic-search”
  • “retrieval-augmented-generation”
  • “ocr”
  • “nanonets”
  • “document-parsing”
  • “coding-assistants”
  • “memory” aliases:
  • “vector database” summary: “A vector store provides semantic similarity retrieval that can be augmented by knowledge graphs for relationship-aware context and OCR for table-to-text extraction.” updated: 2026-04-14 group: applied-ai-workflows

Knowledge Graph

A knowledge graph is a structured representation of entities (nodes) and their relationships (edges), enabling semantic querying and complex relationship-based reasoning. It serves as a foundational component for advanced semantic search and RAG systems.

Core Functionality

• Models knowledge as interconnected entities (e.g., Person, Company, Location) with typed relationships (e.g., WORKS_AT, LOCATED_IN)
• Supports graph traversal queries (e.g., “find all people connected to a company through employment”)
• Integrates with LLMs for natural language query interpretation

Integration in RAG Systems

• Primary role: Provides relationship-aware context beyond semantic similarity
• Light RAG implementation: Combines vector store with knowledge graph for efficient retrieval
• Architecture: Involves chunking documents, extracting nodes and relationships, storing them in both a vector store and a knowledge graph
• Contrast with Graph RAG: Light RAG focuses on simplicity and efficiency

2026 04 14 Build a light RAG system with neo4j

Source Notes

• 2026-04-14: # Private RAG system using notebookLM --- --- https://www.youtube.com/watch?v=aj2FkaaL1co The AI Automators This video demonstrates how to set up and run a fully local, open-source version of Google’s NotebookLM, called InsightsLM. The presenter, Daniel Walsh, highlights t (Private RAG system using notebookLM)