When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

NVIDIA launches AI Blueprint to tackle credit card fraud

NVIDIA AI Blueprint

According to The Nilson Report, global losses from credit card transaction fraud are projected to exceed $403 billion over the next decade. As such, NVIDIA has released an AI Blueprint for financial fraud detection, leveraging the NVIDIA AI platform to help institutions identify and prevent fraudulent credit card transactions.

Unveiled at Money20/20 financial services conference, the Blueprint provides reference code, deployment tools, and an architecture that uses "accelerated data processing and advanced algorithms" to identify subtle anomalies in transaction data. This is said to improve detection accuracy and reduces false positives compared with traditional methods.

Built on the NVIDIA AI Enterprise software suite and accelerated compute, the AI Blueprint is available for customers on Amazon Web Services, with support for Dell Technologies and Hewlett-Packard Enterprise arriving soon. Customers can also access it via NVIDIA partners such as Cloudera, EXL, Infosys and SHI International. By adopting these machine learning tools, businesses may achieve up to a 40 percent improvement in fraud detection accuracy.

According to the tech giant, traditional pipelines often rely on XGBoost for anomaly detection in individual transactions, but they struggle with complex networks of linked accounts and devices. The Blueprint’s core components include NVIDIA RAPIDS, which accelerates data processing to generate features at scale, while graph neural networks (GNNs) is said to help uncover hidden fraud patterns across multiple entities. The GNN embeddings are combined with XGBoost, and real-time inference is optimised by NVIDIA Dynamo-Triton (formerly Triton Inference Server). Both CUDA-X Data Science libraries and Dynamo-Triton are included in NVIDIA AI Enterprise.

Although the Blueprint is currently tailored for credit card transaction fraud, it can be adapted for related use cases such as new account fraud, account takeover and anti-money-laundering screening. Systems integrators, software vendors, and cloud service providers can now integrate the NVIDIA AI Blueprint into their financial services applications to help safeguard customer funds, identities, and accounts against sophisticated fraud schemes.

Report a problem with article
Meta and Constellations Nuclear Energy Agreement
Next Article

Meta announces 20-year nuclear energy partnership to support its data centres

crucial x10 external ssd
Previous Article

Crucial 8TB X10 2100 MB/s USB-C (3.2) external SSD is priced lowest ever at just $440

Join the conversation!

Login or Sign Up to read and post a comment.

0 Comments - Add comment