The Hallway Track
9 tracked signals on fraud-detection.
Wells Fargo built a transformer-based foundation model on tabular transaction data that outperforms XGBoost for fraud detection.
“the model I'm going to show you outperforms XGBoost even in terms of latency, and in fraud detection we get a much better false positive reduction and a higher detection rate”
Apple introduces Trust Insights in iOS 27, a privacy-preserving framework to detect coerced user actions from social engineering scams.
“Authentication confirms who, but not whether they're acting freely.”
Fraud is a property of transaction environment, not the transaction itself
“The signal was not in the string itself, but in the connections. Therefore, fraud is not a property of the transaction, but of the environment in which it occurs.”
California's DFPI built an air-gapped, offline AI system to keep fraud evidence court-defensible.
“generative AI trashed it completely.”
Databricks demonstrates real-time telecom fraud detection using an AI coworker agent architecture
Jumio built a sub-100ms real-time feature store on AWS for fraud detection ML models
Databricks pitches real-time AI fraud detection for federal government benefits programs
Databricks ML pipelines help energy suppliers reclaim 70% of theft exposure worth billions annually
“we've seen energy suppliers reclaim about 70% of their theft exposure by implementing machine learning and AI processes”
AWS integrates SageMaker Canvas ML predictions with QuickSight for no-code fraud detection dashboards