GraphSummit Europe 2024 - The Future of Graph in AI Workloads.pdf
0 likes•111 views
Dr. Jim Webber discusses the evolution of graph technology in AI, highlighting its role in enhancing AI accuracy and driving innovation in various applications like fraud detection and customer service. He emphasizes the importance of knowledge graphs and generative AI as essential components for the future of intelligent applications. The presentation outlines challenges and advancements in production maturity and the automated creation of knowledge graphs using large language models.
GraphSummit Europe 2024 - The Future of Graph in AI Workloads.pdf
- 1. The Future of Graph in AI Workloads Dr. Jim Webber | Chief Scientist, Neo4j Connecting the Dots
- 2. A decade+ of graph innovation GraphConnect London, October 2014
- 3. GraphConnect London, October 2014 A decade+ of graph innovation
- 8. Unlocking the Power of Relationships Faster time-to-market for new models and features Real-time pricing for 100+ markets
- 9. Unlocking the Power of Relationships Knowledge graphs and GenAI on AWS Bedrock Data assets -> business value
- 10. Unlocking the Power of Relationships Enhancing AI accuracy with graph-powered context on AWS Bedrock Driving innovation in fraud detection, risk management, and customer service
- 11. LET’S BE HONEST
- 13. + 😭 = <dependency> <groupId>org.neo4j</groupId> <artifactId>neo4j</artifactId> <version>0.8.0</version> </dependency> for ( Path position : db.traversalDescription() .depthFirst() .relationships( Rels.KNOWS ) .relationships( Rels.LIKES, Direction.INCOMING ) .evaluator( Evaluators.toDepth( 5 ) ) .traverse( node ) ) { output+= position + “n”; }
- 14. The early days
- 15. The early days Neo4j 1.0 2010
- 16. The early days Neo4j 1.0 Neo4j 2.0 2013 2010
- 17. Neo4j 3.0 The early days Neo4j 1.0 Neo4j 2.0 2013 2010 2016
- 18. Neo4j 3.0 Neo4j 4.0 The early days Neo4j 1.0 Neo4j 2.0 2013 2010 2016 2020
- 19. Neo4j 3.0 Neo4j 4.0 Neo4j 5.0 The early days Neo4j 1.0 Neo4j 2.0 2013 2010 2016 2020 2022
- 20. Build Apps, Not Infrastructure
- 21. Making Graph Technology Accessible to Everyone
- 22. AI + Knowledge Graphs: The Next Frontier of Intelligent Apps
- 23. LLMs GenAI go-to-production barrier GPT-3.5 Sparked the GenAI Hockey Stick Moment
- 24. LLMs RAG Raises the Ceiling but it’s Still There RAG GenAI go-to-production barrier Go-to-production use cases unlocked with RAG
- 26. AI Investors Demand ROI as Pressure Mounts Reality: High Hopes, Low Production Maturity
- 27. RAG
- 28. GraphRAG
- 30. GraphRAG Auto-Tuning: Using LLMs to Create Knowledge Graphs for Your Data Automatically creates prompts specific to your domain Adapts quickly to new datasets with less manual work Extracts more relevant entities and relationships for your use case
- 31. Looking to the Future: Agentic AI
- 32. Today we tackled the what and the why. Tomorrow we’ll bend the how!