ai Partners with Southampton University to Enhance Fraud Risk Strategy
Written on 19th, June 2018
The ai Corporation (ai), an FCA approved expert in payments, fraud and risk management, and the University of Southampton (UoS), today announced the launch of a ground breaking research partnership, that aims to further improve fraud risk strategy automation through machine learning.
The project will be funded by Innovate UK, the UK’s innovation agency, via a £100,000 grant over 2 years. The project will commence in the Autumn (2018), with a post doctorate researcher working alongside ai’s research team and Professors from UoS.
ai’s KTP (Knowledge Transfer Partnership) with UoS’s Computer Science and Mathematics Departments will make use of ai’s existing fraud tools, including manual rules and its innovative SmartScore and SmartRule solutions. Increasing understanding of how fraud detection can be automatically and continually optimised, with minimal user interaction on the journey to self-learning.
The joint team will be led by Oliver Tearle, ai’s Head of Research; Professor Mahesan Niranjan – Head of the Information: Signals, Images and Systems Research Group; and Professor Jacek Brodzki – Professor of Mathematics, Head of Applied Topology Group, at UoS.
Oliver Tearle, ai’s Head of Research, says, “Our partnership with UoS is a key step in increasing ai’s ability to further innovate and automate for our clients. The joint research team aims to transfer and embed an innovative capability, based on advanced machine learning approaches, that will automate the time-consuming task of fraud risk strategy management in the finance sector.
“The results of the research will allow fraud and risk managers to take a step back from the continual task of optimising fraud strategy. The machine will be able to work continually in the background, enabling human teams to tackle other challenges in areas that at present can’t be automated.”
Dr Mark Goldspink, CEO of ai, says, “This Innovate UK grant provides endorsement of our payments, fraud and risk management expertise, and reinforces our belief that an innovative approach, using the very latest technologies, is essential to help protect businesses and the public from the latest fraudulent behaviour. Fraud is a global problem and we’re teaming up with some very talented academics to ensure we stay ahead of the fraudsters.”
How ai’s machine-learning technology works
ai’s SmartSuite of machine-learning technology gives banks the power to automatically create effective fraud rules which can be implemented into any fraud platform, including ai’s rules engine RiskNet®. By using machine-learning, banks can automate their fraud prevention, mitigating ACH fraud regardless of how it is perpetrated including account takeover, ‘man in the middle’ fraud and/or social engineering.
ai’s SmartScore® and SmartRule® self-service machine learning based fraud detection systems give users the flexibility of creating an effective, high coverage fraud strategy to detect very specific issues, as well as larger problems. SmartScore®, one of six products within ai’s SmartSuite, creates neural models using artificial intelligence and automated machine-learning techniques, to recognise patterns and trends in fraud. Providing transaction risk scores to be used in conjunction with user-defined rules.
With its unique multi-model capabilities, SmartScore® enables users to create Neural Models specific to a fraud type, customer segment or payment method, including ACH. By constantly refreshing the data available, SmartScore® provides an up to date and accurate risk score based on current trends, ensuring that RiskNet® or any third-party fraud platform users are not reviewing unnecessary alerts.
ai’s current anti-fraud solutions, including its “state of the art” machine learning and artificial intelligence, protect and enrich payments experiences for more than 100 banks, over three million multi-channel merchants and over 300 million consumer cardholders. ai monitors over 25 billion transactions and authorisations every year.