释放 RAG 的力量.pptx

1、Unlocking the Power of RAGKnowledge Graphs:The GPS for AI&LLMs,May 14,2025Data Summit Boston,Presented bySumit PalAtanas Kiryakov,One Stop Shop for Graph AI,Graphwise enables enterprises toconnect the dots across data sourcessurface knowledgepropel decision makingWe leverages graphs to combine domai

2、n knowledge and diverse data into a holistic view that enhances AIs ability to derive meaningful trustworthy insights,Semantic Web CompanyA GRAPHWISE COMPANY,OntotextA GRAPHWISE COMPANY,Graphwise in a Nutshell,200+Happy Clients,20+Years in Business,4.6*Rated by Clients,200+Team Members,*Read more:ht

3、tps:/(4.6 is the average of the GraphDB and PoolParty ratings),infographwise.ai www.graphwise.ai,Presentation Outline,Introduction to Knowledge Graphs Semantic Layers and Multimodal Data FabricsRetrieval Augmented Generation(RAG)and GraphRAGGraphwise Platform for Knowledge Graphs and AIGraph RAG Use

4、 Cases and Killer ApplicationsImproving the Performance and Reducing the Cost of RAGConclusion,Changing Paradigm,OS,Context Window,CPU,Bytes,LMOS,Token,RAM,LLMs,Andrej Karpathy https:/,Graphs are Everywhere,Knowledge Graphs are Everywhere,From Graphs to Knowledge Graphs,From Graphs to Knowledge Grap

5、hs,From Graphs to Knowledge Graphs,How Domain Knowledge and Inference Help?,The benefits:Easily answer an inquiry about“blue sport shoes”In faceted search and In chat(NL question),Graphs Models-LPG,Graphs Models-RDF,Contextualize with Semantics,Add Structure,Context&MeaningParts,relationships,hierar

6、chies,classifications,etc.,Hierarchies,Taxonomies,Ontologies,Contextualize Further with Semantics,Semantic Ontology gives Meaning within a domain“Machine Interpretable”,Every Graph is NOT a Knowledge Graph,Graph of a LAN network is not a KG if there is no schema and semanticsKGs are NOT about search