Inflearn brand logo image
Inflearn brand logo image
AI Development

/

AI Agent Development

graphRAG - Knowledge Graph-based RAG System Implementation with Neo4J (feat. LangChain)

Knowledge Graph-Based RAG System in Neo4J: Next-Generation Search Leveraging Data Relationships Beyond Simple Vector Search Maximize RAG Performance using Graph DB Power!

(5.0) 13 reviews

170 students

RAG
LangChain
neo4j
LLM
DBMS/RDBMS

What you will learn!

  • Building a Knowledge Graph Using the Neo4J Graph Database

  • Applying Knowledge Graph to RAG

  • Implementing a Graph-Based Search System Using LangChain and Neo4J

  • Utilization and integration of various search techniques (Basic search, Expert search, Vector search)

  • Natural Language to Graph Query Conversion using Text2Cypher

Leveraging Graph Databases for Powerful RAG Systems 🪄

General RAG systems rely on simple vector searches, which make it difficult to properly express relationships between information. GraphRAG , which utilizes graph databases, structures complex relationships between data to generate more accurate and contextual responses. Structured knowledge representation and relationship-based search provide more accurate search results .

[Project 2] ETF Financial Product Recommendation - Knowledge Graph Implementation (Using Neo4J Browser)

Improving RAG Performance with Knowledge Graphs 🔧

Structural knowledge representation : Represent information as nodes and relationships to clarify the connectivity between data and systematically structure complex knowledge.

Relationship-based search : Go beyond simple keyword matching or vector similarity to search based on semantic relationships for more accurate results.

Integration of various search methods : Maximize search performance by utilizing keyword-based full-text search and vector-based semantic search together.

Scalable knowledge structure : Easily add and connect new data and relationships to build a continuously expandable knowledge base.

  • Nodes and relationships enable natural data modeling.

  • The Cypher query language enables intuitive graph exploration.

  • Supports powerful graph algorithms and vector indexing .

  • It provides various functions for integration with the LangChain framework.

[RAG implementation using LangChain + Neo4J] https://neo4j.com/labs/genai-ecosystem/langchain/

Features of this course

Step-by-step, practice-oriented learning

Develop a solid understanding and ability to apply concepts of graph databases and RAG integration through a combination of theoretical explanations and immediate hands-on practice.

Learning from various real-life cases

Learn how to build and search knowledge graphs for a variety of domains using real-world data, including movie recommendations, news data, ETF financial products, and legal documents.

Learn about these things

Neo4J Graph Database Basics

Learn about the core concepts of graph databases, such as nodes, relationships, and properties, from setting up Neo4J AuraDB. Acquire basic knowledge required to utilize graph databases, such as the main syntax of the Cypher query language, path navigation, and aggregate functions.

Building a Knowledge Graph

Learn how to transform data from a variety of domains, from CSV structured data to news, ETF financial products, and legal documents, into knowledge graphs. Learn how to design ontology, set constraints, and use Langchain tools for graph transformation.

Implementing a graph-based search technique

Learn how to implement various graph-based search techniques, including basic search, full-text search, semantic search, and Text2Cypher. You will also learn hybrid search methods for implementing enhanced RAG systems.

Things to note before taking the class

Practice environment

  • Operating System and Version (OS): Lecture will be based on MacOS (Linux and Windows users can also practice)

  • Using a virtual environment: Using the uv package manager (conda, poetry, venv users can also practice)

  • Tools used: VS Code, LLM API authentication key required (OpenAI/Google Gemini) *Fees may apply

  • PC Specs: Not applicable

  • Python 3.12 /

    langchain 0.3.23 / langchain-neo4j 0.3.0 / numpy 1.26.4

Learning Materials

  • Providing materials required for practice (lecture notes, practice code, practice data)

  • Check out the list of class notes and class materials to download.

Player Knowledge and Notes

  • Basic Python programming skills

  • Understanding the basic concepts of LangChain

  • No graph database experience required (the course covers the basics)

  • If you have any questions or comments, please feel free to ask.


Recommended for
these people!

Who is this course right for?

  • Developer looking to improve the performance and accuracy of RAG systems

  • For those who want to effectively handle data with complex relationships

  • Someone who wants to build a structured knowledge search system beyond simple vector search

  • Those who want to apply Knowledge Graphs to AI applications

Need to know before starting?

  • Python

  • LangChain

  • RAG

Hello
This is

12,730

Students

386

Reviews

117

Answers

4.9

Rating

7

Courses

안녕하세요. 저는 파이썬을 활용한 데이터 분석 및 인공지능 서비스 개발 실무를 하고 있습니다. 관심 있는 주제를 찾아서 공부하고 그 내용들을 많은 분들과 공유하기 위해 꾸준하게 책을 집필하고 인공지능 강의를 진행해 오고 있습니다.

 

[이력]

현) 핀테크 스타트업 CEO

전) 데이콘 CDO

전) 인덕대학교 컴퓨터소프트웨어학과 겸임교수

Kaggle Competitin Expert, 빅데이터 분석기사

 

[강의]

NCS 등록강사 (인공지능)

SBA 서울경제진흥원 새싹(SeSAC) 캠퍼스 SW 교육 ‘우수 파트너 선정’ (Python을 활용한 AI 모델 개발)

금융보안원, 한국전자정보통신산업진흥회, 한국디스플레이산업협회, 대구디지털산업진흥원 등 강의

서울대, 부산대, 경희대, 한국외대 등 국내 주요 대학 및 국내 기업체 교육 경험

  

[집필]

 

[유튜브] 판다스 스튜디오 : https://youtube.com/@pandas-data-studio?si=XoLVQzJ9mmdFJQHU

Curriculum

All

58 lectures ∙ (7hr 4min)

Course Materials:

Lecture resources
Published: 
Last updated: 

Reviews

All

13 reviews

5.0

13 reviews

  • sketchesfancy9795님의 프로필 이미지
    sketchesfancy9795

    Reviews 3

    Average Rating 5.0

    5

    31% enrolled

    I want to quickly try implementing in my DB.

    • jyuri0018155님의 프로필 이미지
      jyuri0018155

      Reviews 1

      Average Rating 5.0

      5

      31% enrolled

      • kmkang2281님의 프로필 이미지
        kmkang2281

        Reviews 6

        Average Rating 4.8

        5

        31% enrolled

        Thank you sincerely.

        • hshin25375075님의 프로필 이미지
          hshin25375075

          Reviews 2

          Average Rating 3.0

          5

          31% enrolled

          • adastra01190030님의 프로필 이미지
            adastra01190030

            Reviews 1

            Average Rating 5.0

            5

            100% enrolled

            After using vectorRAG, I wanted to learn about graphRAG, and fortunately a course became available, so I took it. It was beneficial as I could follow along and learn through the Kind practical files, from explanations of basic concepts, data processing using real data, various query search methods... up to Hybrid RAG! There was something I actually wanted to implement, and after taking this course, I think I can immediately start using the knowledge I learned. The instructor's tone was calm and flowed smoothly, so even when listening at 1.7x speed, it wasn't bothersome and was easy to understand, allowing me to complete the course quickly! I enjoyed the course very much, instructor. I look forward to your next course as well!

            pdstudio's other courses

            Check out other courses by the instructor!

            Similar courses

            Explore other courses in the same field!