Explore real-time fraud analysis in a hybrid cloud environment through this conference talk. Discover how open source and closed source technologies can be combined to build and train AI models for fraud detection in the cloud, then deploy them for real-time analysis on mainframes. Learn about using TensorFlow for model training, persisting models in ONNX format, and leveraging IBM's Telum AI Accelerator for high-scale, in-transaction inference. Gain insights into overcoming challenges in AI adoption, securing data in compliance with financial industry regulations, and applying this hybrid approach to various use cases beyond fraud detection. Understand the AI lifecycle, from data sourcing to deployment, and explore the role of open source technologies in solving real-world problems across industries in a hybrid multi-cloud environment.
Open Source Powered Real-Time Fraud Analysis in a Hybrid Cloud Environment
Overview
Syllabus
Introduction
Agenda
AI Impacts Every Industry
Challenges in AI Adoption
Ideal AI Stage
Real World Use Cases
Other Use Cases
Customer Use Case
How Open Source Helps
AI Life Cycle
AI Modeling
Data Source
Deploy
TensorFlow Model
Deep Learning Compiler
Swagger API
Data Security
Data Protection
IBM Cloud
Model Deployment
Machine Learning Models
Open Source Models
Cloud Object Storage
Data Mapping
Testing
Summary
Open Source in 10 years
Taught by
Linux Foundation
