Cyber Security projects include an AI-enabled stress-busting platform and Identity Security for IoT

Our digitally connected world is hugely enabling. But its connectedness leaves us all potentially vulnerable to cyber security attacks.  

In the second of two articles we spotlight some of the projects under development via the Cyber Security Academic Accelerator Programme (CyberASAP) – an initiative which supports UK academic teams to commercialise their cyber security ideas. 

The programme is funded by the Department for Digital Culture Media and Sport and delivered by KTN in partnership with Innovate UK. This work is part of the government’s £1.9bn national cyber security strategy and supports the ambition to make the UK the safest place to live and work online

Since the programme started in 2016, participating projects have attracted over £6million in further funding. 

Half of the projects from this year’s cohorts were showcased in this article. The remaining seven are detailed below.  All 14 teams are building towards developing a Proof of Concept which will be showcased to potential investors and industry partners at a Demo Day in February. If you are interested in participating, please register here. 

Edinburgh Napier University – Memcrypt 

Memcrypt protects and recovers confidential data from ransomware attacks. 

Ransomware is a form of malicious software designed to block user access to files by encryption until a sum of money is paid. It is a growing global problem with estimated costs of $169 billion in 2020. Even without paying a ransom, costs include lost business, recovery time, third-party  remediation services, and reputational loss. 

Existing methods for combating ransomware do not enable the user to quickly recover from an attack, when ransomware has succeeded in starting to encrypt user data. Memcrypt’s core innovation will discover active ransomware keys and related artefacts and enable almost immediate data recovery.  The proposed prototype, Memcrypt Triage, will provide an out-of-the box ransomware triage tool for incident response. The tool can be used to scan a target which  has been affected by ransomware and will collect information on files affected by a ransomware  attack. This will aid law enforcement in the collection of digital forensic evidence and also help the development of our technology in ransomware key detection and data recovery. For more information on this project please contact us.

Imperial College London – WhatML: Watermarking Machine Learning Models

WhatML protects the value and the intellectual property of machine learning models. 

Training machine learning models requires acquisition and processing of huge amounts of data, and significant investment both in computing infrastructure and in the skills required to train the models effectively. As a result, these machine learning models are hugely valuable assets that need to be protected. In many cases, these models are exposed, as they need to be deployed in the customers’ facilities or in the final product. Other business models rely on outsourcing the training tasks to external suppliers or the use of marketplaces to buy and sell models specific to a task. All these forms of monetisation cannot be used without adequately protecting the models against theft, illegal copying and use beyond contractual terms. To address these shortcomings, WhatML is a solution to protect the intellectual property of machine learning models through watermarking which is resistant to different model transformations and enables verification of the models’ ownership or provenance. A novel and innovative solution as there are currently no commercial products to protect the intellectual property of machine learning models. For more information about this project contact cyberasap@iuk.ktn-uk.org.

University of Kent – #ID Security for IoT 

Secure Device Identity to power the future of the Internet of Things.  

IoT devices are becoming ubiquitous with early applications ranging from smart-meters in the  home through industrial control systems in factories to automated cars. The market potential for  such systems is very high. However, this risk of compromise to the security of such systems makes them highly vulnerable, as existing security systems are inadequate for such applications. The ability to provide a high performance protection system follows from its capacity to derive device identifying #ID’s directly from the operating characteristics of IoT devices. This contrasts to the  traditional approach where, for identification, a device either stores an identifier within it or typically submits a sample during a process called enrolment, and a digital representation of the sample is then stored as a template. Significantly, our proposed system does not store any templates or copies of the #ID’s and therefore the opportunity for system breach via potential compromise is completely eliminated. This provides a disruptive technology with the capacity for enormous impact. For more information about this project contact cyberasap@iuk.ktn-uk.org.

Lancaster University – Developer Security Essentials 

A non-profit helping consultants make the UK’s 400,000 developers better at security.  

Developer Security Essentials offers a cost-effective way to help any software development team improve the security and privacy of their code. It promotes cyber security and privacy as a  business asset, makes them comprehensible to developers, and introduces the techniques  needed to implement them correctly. 

While a majority of UK software development teams still use at most one security technique, new laws mean that their organisations are penalised for security breaches, and cloud computing and DevOps mean that security must now be entirely in their code. Developer Security Essentials provides a solution. It is a half day package of structured workshops to motivate and empower developers to produce secure code, designed for non-specialist consultants and trainers to present. For more information about this project contact cyberasap@iuk.ktn-uk.org.

Middlesex University – Linux 

A Security assessment tool for Linux systems based on the MITRE Framework. 

Linux is the predominant operating system for Internet Services and the National Critical  Infrastructure. This makes these systems appealing targets for hackers. Unfortunately, Linux’s  built-in security access controls (SELinux and Apparmor) are underutilised due to their complexity; and the lack of support by third party security assessment tools makes these systems even more vulnerable. Therefore, we propose SALMAC, a unique security assessment tool focusing on hardening SELinux and Apparmor in order to reduce the attack surface and stop misconfiguration-based attacks in the Linux environment. 

Designed using a well-maintained threat model SALMAC gauges systems’ access controls against targeted attacks. Whilst simulating attacks, it pulls system logs from the targeted machines and links them to the launched attacks. This Event-Log pairing forms the basis for effective threat hunting as well as being used to build the intelligence for automated detection of future attacks, helping support the NCSC’s vision of “Making the UK the safest place to live and do business online”. For more information about this project contact cyberasap@iuk.ktn-uk.org.

University of Southampton – CyberHelper 

CyberHelper is an innovative tool that efficiently runs your cyberattacks’ investigations.  

Our society is facing an increase of interconnectivity that comes with its own cybersecurity  challenges. Security analysts analyze the data related to threats or cyberattacks, in order to put in place fast and efficient countermeasures. Frequently, the analysts find themselves  overwhelmed with data to be analyzed and with tools that require a high level of expertise to be used. Both factors, together with the time-pressure on the analysts, create a negative impact on the analysis and aggravate the already existing human-error element. CyberHelper is a software solution for the analysis of threats and cyberattacks. This innovative tool combines network security, AI, knowledge representation and decision making, to guide security analysts in the analysis of cyberattacks. CyberHelper works with different types of information and increases the efficiency of the investigation by providing the analyst with AI-driven insights and guidance, thus reducing the investigation time. For more information about this project contact cyberasap@iuk.ktn-uk.org.

University of Wolverhampton – CyberMIND 

CyberMIND is an AI-based platform helping Cybersecurity professionals to detect, predict, and manage stress.  

Cyber professionals, like front line soldiers in a warzone, protect our organisations and critical infrastructure. But due to an increase in the number of sophisticated attacks, shortage of skilled staff and overwhelming workloads nearly 50%* of them are suffering from mental health issues. 40%** may leave by the end of year due to stress. So, what will that cost? And how secure are we? CyberMIND helps predict and manage the mental health and wellbeing of cyber teams. The result? Improved performance and productivity; reduced cyber risk; plus compliance with the Duty of Care to assess employee stress levels – and all with an ROI in just a few days? For more information about this project contact cyberasap@iuk.ktn-uk.org.

*Source: Nominet Report on Stress February 2020

**Source: Ciisec Report 

Find out more about CyberASAP

CyberASAP is in its fourth year. Led by KTN’s Emma Fadlon and Robin Kennedy the programme delivers a comprehensive series of workshops presented by KTN and its extensive network of expert mentors and collaborators. The programme helps upskill the selected academic teams during their commercialisation journey and is designed to de-risk their path to market and encourage their innovative ideas to flourish. 

Find out more about CyberASAP here.