The two-year project, co-financed by Innovate UK, will develop and embed an intelligent data driven and knowledge-based fraud prevention and detection service to support insurance claim handling. It will use modern machine learning, data analytics techniques, semantic technologies, intelligent modelling methodologies and decision support systems.
The academic team supporting the two-year project were Jian-Bo Yang, Professor of Decision and System Sciences and Director of the Decision and Cognitive Sciences Research Centre (DCSRC), and Dong Ling Xu, Professor of Decision Science and Support Systems.
Explained Prof Yang: “Fraud is a major enemy of insurance companies, with huge costs not just to the industry but also to consumers via increased premiums. At the moment the industry tries to prevent fraud by looking for and detecting suspicious patterns but criminals are very clever and always trying to stay one step ahead, so attempting to identify new patterns as soon as they arise is key.”
Analysis
There are three stages to the KTP. Firstly academics will examine the current system that Kennedys use and assess whether it can be improved via data mining and knowledge-based modelling.
Secondly, they will propose a new algorithm and software system using the knowledge base and aggregated databases that Kennedys has developed. And finally they will seek to demonstrate the potential of the new system for the rest of the world.
Added Prof Yang: “We are confident that we can help Kennedys improve their current system. The key is to try and develop a hybrid system where you can use both big data and human knowledge in deep learning to tackle the problem, which we call transparent machine learning.
“In this way you can explain exactly why you reach your decisions. It is evidence-based, transparent decision-making. Ultimately you have to be able to explain to stakeholders what you are doing. It is hard for people to trust a black box that does not explain how it reaches a conclusion. This KTP could potentially be a great impact case study for us if it can be proven to reduce fraud across the insurance industry.”
Global reach
Kennedys is an international law firm with expertise in litigation and dispute resolution, particularly in the insurance/reinsurance and liability industries. Worldwide it has 36 offices in the UK and Europe, Middle East, Asia Pacific and the Americas, and acts for insurers, reinsurers, Lloyd’s Syndicates, public bodies and corporates.
This is the second KTP to be awarded to AMBS with a focus on legal technology. It is also one of a number of KTPs in the area of data analytics and decision support being pursued by DCSRC and AMBS.
Prof Yang said the same principles around transparent machine learning, artificial intelligence and evidence-based decision support can be applied to many other sectors such as law, healthcare, financial technology (FinTech) and engineering systems.
“For instance we have also been talking to hospitals and health research organisations such as Manchester Institute of Population Health and Beijing Xuanwu Hospital about more accurate diagnosis of complicated disease such as cancer, asthma, sepsis and trauma. The key is that the conclusions you reach are interpretable, data driven and based on expert knowledge as well,” he added.
Productivity
Meanwhile, a report published this month by the CBI setting out new ways to tackle variations in productivity between UK firms, calls on Innovate UK to expand the KTP model, in particular giving businesses more opportunities to work with business schools.
*If your business would like to explore potential projects that may benefit from specific academic expertise contact joanne.summers@manchester.ac.uk who can explain KTPs in more detail. The KTP programme currently has a national success rate of 86% and additional funding has been awarded to the KTP programme via the Industrial Strategy Challenge Fund, enabling 200 new KTPs by the end of the current financial year.
