A single cyber breach, depending on the country and industry, has been estimated to cost on average $3.86 million, rising to an average of $8.64 million for the USA.
Threats to cyber-security are a global problem, affecting business, governments and society. Maintaining secure environments is hugely important across national infrastructures, and to secure these systems and safeguard the privacy, wealth and wellbeing of all members of society, statistical and machine learning techniques have an important role to play.
The WannaCry ransomeware attack
In May 2017, the WannaCry ransomware attack indiscriminately targeted older Microsoft Windows operating systems that were not patched to close a known exploit. This included around 200,000 computers across 150 countries, causing financial losses reaching approximately $4bn and causing the National Health Service in England and Scotland to turn away non-critical emergencies and divert ambulances as it tried to deal with the impact of the attack.
Responding to the threat
As part of work that has been ongoing since 2010 researchers from the Department of Mathematics, led by Nick Heard and Niall Adams, have developed statistical methods and data science techniques, which have helped strengthen cyber-security across systems used by hundreds of millions of people globally.
This innovation of statistical models for network host behaviours and traffic patterns observed within an enterprise computer network has far reaching applications and benefits. The research group has collaborated with Los Alamos National Laboratory, Microsoft and EY, and these collaborations have led to joint research publications and patents, which have subsequently been adopted in cyber-security software benefiting users worldwide with improved security and safety.
Meet the mathematicians
Lead researchers Professors Niall Adams and Nick Heard tell us more about their work:
Why has this research had such effective societal impact?
In our increasingly connected world, cyber-security is a global concern which affects all aspects of modern life. Government departments, multinational corporations, small businesses, schools, hospitals and financial service providers have all been affected by high-profile attacks, and the potential losses and disruptions at stake have few limits. With human error perhaps the largest source of vulnerability within any computer network, cyber-security can never be fully guaranteed.
How does maths underpin this work?
Interdisciplinary work in this area is crucial. Huge data volumes can be captured from an enterprise computer network, and real-time processing of so much data seems a formidable challenge. Bespoke statistical tools are required for analysing these data streams, filtering, modelling and then performing anomaly detection to identify statistical outliers. Domain expertise is vital for advising which discoveries are uninteresting (most of them), and those (few) which are interesting.
What are the most challenging and rewarding aspects of contributing to this work?
The research challenges are interesting for their variety. The same streams can be viewed through a range of analytical perspectives, from macro-level analyses using graph theoretic techniques for understanding entire network structures, down to micro-level modelling, examining the patterns and behaviours in connectivity on just one edge in that larger network graph.
How do you see this area of research evolving?
Over the next few years, the impacts of statistical cyber-security methods will become more apparent. A global survey suggests cyber-security is the top investment priority in business, and artificial intelligence is likely to be an integral characteristic of the security tool set evolution. But this is a double-edged sword, with increasingly sophisticated malicious AI techniques also likely to emerge.
Any tips for students who are interested in embarking on a career in cyber security?
There are excellent opportunities in both academia and business, and there is a diverse range of areas of expertise which can be relevant. Many students might not be aware that this is an emerging but important application area for statistics.
View other impact case studies
Find out more about some of the research happening in the Department of Mathematics and how our academics are having real-world impact.
