Demo: Real-time Machine Learning Execution
This demo shows how a farm of ML model clones can be fed events arriving live from multiple concurrent users, and makes decisions then and there. The ability to make decisions in real-time allows user behavior to be influenced and for the user experience to improve. For fraud, this might mean improved detection of suspicious transactions. For trading, this might mean predicting market movements. For product application forms, more rapid decisions result in higher take-up.
The speaker was Neil Stevenson.