Financial Services

American Express Prevents Fraud and Foils Cybercrime With NVIDIA AI Solutions

Objective

American Express uses AI for ultra low latency fraud detection in credit card transactions.

Customer

American Express

Use Case

Fraud Detection

Technology

NVIDIA DGX^™, NVIDIA TensorRT^™, NVIDIA Triton^™ Inference Server

Using AI to Combat Financial Fraud

Credit and bank cards are a major target for fraud. To thwart fraudulent activity, American Express, which boasts over 115 million active credit cards and handles more than eight billion transactions a year, leverages deep learning through the NVIDIA GPU computing platform.

NVIDIA Solutions

American Express uses advanced, long short-term memory (LSTM) models to detect anomalous patterns in transactions. Models are trained on NVIDIA DGX™ for faster time-to-market, optimized with NVIDIA TensorRT™, and served with NVIDIA Triton™ Inference Server for high performance in production.

American Express Results

American Express implemented this enhanced, real-time fraud detection system to improve accuracy, and better protect customers and merchants. The system operates within a tight two-millisecond latency requirement and delivers a 50X improvement over a CPU-based configuration, which failed to meet their goal.

A GPU-accelerated LSTM deep neural network, combined with their long-standing gradient boosting machine (GBM) model—used for regression and classification—has improved fraud detection accuracy in specific segments.

Results

50X the throughput on GPUs vs. CPUs

About American Express

American Express is a leading issuer of personal, small business, and corporate credit cards. The company’s travel-related offerings include traveler’s checks, credit cards, corporate and personal travel planning services, tour packages, and agencies for hotel and car-rental reservations. American Express handles more than eight billion transactions a year.

“Our fraud algorithms monitor, in real-time, every American Express transaction around the world for more than $1.2 trillion spent annually, and we generate fraud decisions in mere milliseconds. Having our card members’ and merchants’ backs is our top priority, so keeping our fraud rates low is key to achieving that goal. Especially in this environment, our customers need us now more than ever, so we’re supporting them with best-in-class protection and servicing.”

Vice President of Machine Learning and Data Science

American Express