2880- AI Use Cases on IBM Power10.pdf

1、AI Use Cases on IBM Power10 2880Youssef LargouCEO Power Maroc and CTO rKube.IOIBM Champion and IBM Redbooks Platinum AuthorErik RottsolkHead of Partnerships,Sr.Account Executive,Trovares IncOctober 21October 21-24,2024 Mandalay Bay 24,2024 Mandalay Bay Las VegasLas VegasAgenda2IBM TechXchange/2024 I

2、BM CorporationIntroductionTraditional AI vs Generative AIAI Efficiency with IBM Power 10Clients use casesDemo:Graph Analytics Revolution with Trovares AIQ&A010203040506Agenda3IBM TechXchange/2024 IBM CorporationIntroductionTraditional AI vs Generative AIAI Efficiency with IBM Power 10Clients use cas

3、esDemo:Graph Analytics Revolution with Trovares AIQ&A010203040506Enterprise considerations:Traditional AI is still critical,Gen AI unlocks new possibilities Foundation models majorly trained with unlabeled dataTraditional AI(ML/DL)models majorly trained with labeled dataWorks well for general tasks

4、and can improve for specific tasks with relatively less trainingA tuned model can be very efficient for the specific task it was designed forTrained for no specific taskTrained for a specific taskTrained on very big data sets(100s of TBs)Trained on proper,large data sets(100s of GBs)Generative AIGen

5、erative AITraditional AITraditional AIIBM TechXchange/2024 IBM CorporationIBMs POV for AI-ready IT InfrastructureReliable performanceScale AI inferencing for complex tasks like generative AIHybrid flexibilityCreate AI workflows based on where your data and applications resideSecured insights Offer s

6、ecurity and data protection to promote trust and support compliance demandsIBM TechXchange/2024 IBM CorporationAgenda6IBM TechXchange/2024 IBM CorporationIntroductionTraditional AI vs Generative AIAI Efficiency with IBM Power 10Clients use casesDemo:Graph Analytics Revolution with Trovares AIQ&A0102