USI Cybersecurity Risk Project wins AXA Post-Doctoral Fellowship

(image courtesy of Forbes, via Flickr)
(image courtesy of Forbes, via Flickr)

Institutional Communication Service

13 October 2016

Overall worldwide losses from card fraud are expected to double by 2020. This Cybersecurity research project uses advanced statistical techniques to minimize the time required to detect a fraud while maximizing the accuracy of the detection. The proposal achieved high marks from the AXA Research Fund due to the high quality proposal, innovative approach, and access to live big data from a leading credit card services company.

Consumers increasingly make their purchases by credit, debit and money transfer systems such as PayPal, Skrill, and Google Wallet, firmly embracing eCommerce as an integral part of their lives. This revolutionary change in the marketplace has resulted in innovative purchase platforms leading to positive customer experiences but has also caused greater exposure to Cyberfraud. As financial institutions attempt to contain losses by blocking cards when a potential theft is detected, a further risk is introduced: unnecessary card blocks cause frustration with customers and can result in negative feelings towards the financial institution, or reputational risk.

The project Controlling Cybersecurity Risk: Fast Fraud Detection using sequential and optimal stopping techniques approaches the problem from both sides simultaneously, minimizing the time required to detect a fraud while maximizing the accuracy of the detection. This results in lower losses for the financial institution and fewer false card blocks and thus a better experience for the customer, an overall reduction of risk for all parties. The project develops a double threshold model to address both risks concurrently, issuing a warning if the probability of a fraud reaches a certain level, and issuing a block when such probability goes over an upper threshold. To create this model, optimal stopping theory is applied to pre and post fraud spending patterns to produce an accurate prediction model.

The global importance of this research was confirmed when it won the AXA Research Fund’s Post-Doctoral Fellowship, awarded annually to 25 projects worldwide that address the theme of societal risk. This is the first grant that USI has won from the international insurance giant. The project will be conducted at USI by Dr. Bruno Buonaguidi, under the supervision of Prof. Antonietta Mira. A major Swiss credit card services company will provide the data, and the calculations will be made at the Swiss National Supercomputing Centre, CSCS.

Read more about the project
For information, contact: [email protected]