Swarm Learning and Graph Analytics Create Another Quantum Leap in Fraud Detection
Fraud is a serious concern for nearly everyone, especially businesses operating in financial services. At the end of 2021, the annual Nilson Report predicted that credit card fraud will cost the credit card industry collectively $408.5 billion over the next decade.
It’s a lucrative target, following the famous, but apocryphal, Willie Sutton quote. Sutton, when purportedly asked by a reporter why he robbed banks, was reported to answer, “because that’s where the money is.”
While Sutton later denied ever having made that statement, it’s all too true. Criminals will focus their attention on where the money is. With payment card volumes estimated in 2030 to reach nearly $80 trillion, the opportunity is huge. In response, targeted companies like credit card providers are keen to find solutions that will detect fraudulent activity and mitigate, if not entirely prevent, losses from fraud.
At TigerGraph, we’re taking a quantum leap in fraud detection with swarm learning, powered by HPE.
Swarm Learning: Decentralized Machine Learning
HPE’s Swarm Learning solution, developed by Hewlett Packard Labs, is a decentralized machine learning framework that allows insights to be shared amongst distributed systems in a secure manner. This allows for systems using the framework, like fraud detection solutions, to learn and respond to broader datasets more efficiently and elegantly than using traditional methodologies.
In HPE’s press release, Justin Hotard, executive vice president and general manager, HPC & AI, at HPE, describes swarm learning as follows:
“Swarm learning is a new, powerful approach to AI that has already made progress in addressing global challenges such as advancing patient healthcare and improving anomaly detection that aid efforts in fraud detection and predictive maintenance. HPE is contributing to the swarm learning movement in a meaningful way by delivering an enterprise-class solution that uniquely enables organizations to collaborate, innovate, and accelerate the power of AI models, while preserving each organization’s ethics, data privacy, and governance standards.”
Fraud Detection Using the Swarm
If you stop and think about the ‘edge’ in the credit card payment industry, the idea of a swarm isn’t far off. Consider all of the point of sale card readers, ATM machines, mobile device card swipe systems, and online marketplaces in existence today. Each of those connects to some kind of processing command and control, security, and record-keeping system that you can’t see.
In 2018, there were nearly 369 billion credit card transactions reported globally. That means, on average, over one billion transactions occurred … every day. It’s safe to assume that just a few of those transactions were initiated by fraudulent actors.
And we’re only talking about credit card transactions here!
By combining our fully distributed, native parallel graph database with HPE Swarm Learning, credit card companies can rapidly detect unusual credit card transactions. The accuracy of the solution will continue to increase, making it even more efficient, as it learns from the massive number of transactions running through it, all handled securely at the edge thanks to the swarm learning framework.
HPE’s Swarm Learning announcement represents a quantum leap in the fraud detection game, one of many that we are making at TigerGraph, and certainly one we think will make a big difference for the financial services industry and beyond. Learn more about HPE Swarm Learning in the company’s recent press release.