BAE Systems Digital Intelligence have partnered with the University of Nottingham on examining how unsupervised machine learning can be used in the discovery of previously undefined risk, known as “missed risk” in the context of financial crime.
Teams from both organisations will work together using BAE Systems existing Financial Testing Service (FTS) data for digging out complex signals which are difficult to identify with current traditional rules. Grounded in customer needs, the team will initially focus on three use cases: detecting human trafficking, crypto risk assessment and uncovering shell companies. Once models are developed, BAE Systems will engage with customers to share the insight and results in order to explore routes for exploitation together.
“In the AML domain, unsupervised machine learning is in its infancy. As a forward-thinking company, we want to stay ahead of the curve and provide our customers additional security against unforeseen and unusual events,” said Martin Barber, VP, People & Capabilities for BAE Systems Digital Intelligence and project sponsor.
“By partnering with the University of Nottingham Malaysia, academia and analysts will work side by side to push the boundaries of machine learning for enhanced AML solutions,” he added.
“This partnership will use the university’s experience in other domains and learnings from known research to navigate through the research area and bring the requisite rigour to our analysis,” said Dr. Chen Zhi Yuan, Associate Professor, Head of School of Computer Science, University of Nottingham Malaysia.
“This agreement will allow us to expand our understanding and insight on unsupervised machine learning and apply it to realistic financial crime dataset to form an initial assessment,” Dr. Chen added.
“New insights from machine learning could assist our customers by spotting unknown risky events in the data and informing better predictable solutions for the future,” said Dr. Nicholson.
“Our customers want to dig deeper into their data and this partnership allows us to push the boundary on that,” he concluded.
