GraphRAG and Knowledge Graphs Exploring AI's Future
0 likes•348 views
The document discusses the evolution and future of knowledge graphs, particularly in the context of AI and data relationships. It emphasizes the importance of semantic search and how knowledge graphs can enhance search results by linking ideas through graph exploration. Additionally, it introduces tools like Graphrag and LLM knowledge graph builders that improve the accuracy and explainability of AI applications.
![11
Cypher: Pattern Based
MATCH (p:Employee {employee_ID: 1})-[r:WORKS_AT*..3]-(c:Company)
RETURN c.name as company, count(*) as strength
ORDER BY strength DESC
Node
Pattern
Relationship
Pattern](https://image.slidesharecdn.com/cdl2024-jesusbarrasa-241217113740-449d1efe/85/GraphRAG-and-Knowledge-Graphs-Exploring-AI-s-Future-11-320.jpg)
GraphRAG and Knowledge Graphs Exploring AI's Future
- 1. GraphRAG and Knowledge Graphs Exploring AI's Future Dr. Jesús Barrasa Head Solutions Architecture-Neo4j
- 2. Neo4j Inc. All rights reserved 2024 2 But first a quick look at the past… CDL2016
- 3. Neo4j Inc. All rights reserved 2024 3 https://www.w3.org/DesignIssues/RDB-RDF.html And even further back… 2004 😱
- 4. Neo4j Inc. All rights reserved 2024 4 https://github.com/jbarrasa/goingmeta/tree/main/session27 http://goingmeta.live
- 5. The Graph The shape of data Neo4j Inc. All rights reserved 2024 5
- 6. Explore the connections in your data to unlock deeper insights
- 7. Neo4j Inc. All rights reserved 2023 Neo4j Inc. All rights reserved 2024 A ring… or a money laundering scheme
- 8. Neo4j Inc. All rights reserved 2023 Neo4j Inc. All rights reserved 2024 r a critical device in a network or a critical device in a network A star…
- 9. Neo4j Inc. All rights reserved 2023 Neo4j Inc. All rights reserved 2024 or a community of people A cloud…
- 10. 10 The Property Graph: Simply Powerful Employee City Company Nodes represent objects (nouns) Relationships are directional Relationships connect nodes are represent actions (verbs) Relationships can have properties (name/value pairs) Nodes can have properties (name/value pairs) name: Amy Peters date_of_birth: 1984-03-01 employee_ID: 1 :HAS_CEO start_date: 2008-01-20 :LOCATED_IN
- 11. 11 Cypher: Pattern Based MATCH (p:Employee {employee_ID: 1})-[r:WORKS_AT*..3]-(c:Company) RETURN c.name as company, count(*) as strength ORDER BY strength DESC Node Pattern Relationship Pattern
- 12. The Knowledge The semantic layer Neo4j Inc. All rights reserved 2024 12
- 13. Two “types” of semantics Neo4j Inc. All rights reserved 2024 13
- 14. Two “types” of semantics Neo4j Inc. All rights reserved 2024 14
- 15. Two “types” of semantics Neo4j Inc. All rights reserved 2024 15
- 16. Neo4j Inc. All rights reserved 2024 16
- 17. A Knowledge Graph captures key enterprise knowledge in the form of entities and relationships between them. Some nodes in the graph have properties with NL text Neo4j Inc. All rights reserved 2024 17
- 18. These property values get embedded (transformed into a numeric representation) and added to a vector index to enable vector-based semantic search. Neo4j Inc. All rights reserved 2024 18
- 19. A semantic search on the vector index returns the k approximate nearest neighbours to the search concept (word, question, image, etc) Neo4j Inc. All rights reserved 2024 19
- 20. Each result from the vector search is “dereferenced“ to get the corresponding node in the graph and a subsequent graph exploration finds semantically related elements that enrich and augment the final search result. Neo4j Inc. All rights reserved 2024 20
- 21. GraphRAG Neo4j Inc. All rights reserved 2024 21
- 22. Neo4j Inc. All rights reserved 2024 22 Grounding LLMs with KG: GraphRAG
- 23. Real RAG is agentic Neo4j Inc. All rights reserved 2024 23 https://graphrag.com/
- 24. Graph Shapes + Retrievers Neo4j Inc. All rights reserved 2024 24 Lexical Graph with Hypothetical Questions Lexical Graph with Hierarchical Structure Lexical Graph with Extracted Entities LexiGraph w ExtEnt and community summaries https://graphrag.com/
- 25. Benefits of GraphRAG Neo4j Inc. All rights reserved 2024 25 1. Much Higher Accuracy 2. Easier Development 3. Explainability and governance
- 26. LLM Knowledge Graph Builder Neo4j Inc. All rights reserved 2024 26 https://neo4j.com/labs/genai-ecosystem/llm-graph-builder/ https://github.com/neo4j-labs/llm-graph-builder
- 27. GraphRAG Python Package Neo4j Inc. All rights reserved 2024 27 https://neo4j.com/blog/graphrag-python-package/ https://neo4j.com/docs/neo4j-graphrag-python/
- 28. Thank you! [email protected] Neo4j Inc. All rights reserved 2024 28