2023 Funded Projects

Project Description

A custom cybersecurity tool designed for flexible managed security services which involves building a new detection engine using data lake, machine learning, and Artificial Intelligence (AI) to support any data added to the lake for potential correlation to a cyber event. Project Atlas resolves many limitations of cloud and off the shelf products that restrict our managed security services to match the capability of the platform. Current solutions are also designed for end users as opposed to managed service providers, further restricting our ability to provide end-to-end security services. Atlas also allows organizations to consolidate their security vendors into this solution which is financially beneficial. Project Atlas creates a more cost effective and comprehensive security service with less false-positives.

Project Description

Data backup systems use data deduplication techniques, to reduce the stored data size, support more users, and reduce service costs. Data deduplication is a widely used technique by modern cloud-based storage systems such as Acronis Cyber Protect and Dropbox. State-of-the-art techniques for deduplication require storing the data without encryption. Consequently, clients must have complete trust in the service provider and its cybersecurity protections. This trust model is unacceptable to clients and businesses with regulatory and privacy constraints. In this project, the research team will explore building Backup with Confidential Dedup (BCD), a backup system that performs data deduplication while keeping the data encrypted at the service provider.

Project Description

The shift towards 5G networks will be accompanied with the coexistence of previous generations (e.g., LTE 4G, whichexposes 5G networks to security risks related to vulnerabilities of its predecessors due to the inter-working modes and the threats related to the cross-protocol attacks. 5G adoption of new enabling technologies (such as NFV, SDN, and network slicing) increases its attack surface. This project studies 5G security concerns by developing intelligent anomaly detection, mitigation and prevention mechanisms using Machine Learning and Artificial Intelligence solutions. Thereby providing mobile operators with automated, closed-loop control security mechanisms that secure their network against attacks that hinders their network availability, and ensure 5G clients the services they desire.

Project Description

The project involves the development and use of artificial intelligence in vulnerability management including the process of identifying and mitigating vulnerabilities in humans, computer systems and configurations, networks, and software applications. The traditional approach of vulnerability scanning, and management becomes inefficient because it overlooks risk factors that are usually indicators of complex attack vectors such as those user related weaknesses so the use of AI and machine learning systems and vulnerability management will not only make the process faster, but it will also make it more efficient. AI enabled vulnerability management can help facilitate the scoring and ranking of threats, which means less time spent looking for clues and less money spent on vulnerability scanning.

Project Description

Different people behave in different ways while visiting online websites or while engaging in the metaverse. This forms a user’s behavioural fingerprint, which can be used for benign or malicious purposes. This project collects various datasets with behavioural data from both real-world websites and the metaverse to determine which features of the collected data are useful for creating individual user fingerprints from these datasets. It will establish the potential of using these fingerprints for several purposes, including, user profiling, user impersonation, fraud detection, and fraud prevention. The goal is to detect and defend against fraudulent activities such as with credit cards.

Project Description

IoT device vulnerabilities are increasing. Recently the adoption of fuzzing techniques for automated vulnerability detection in IoT firmware has evolved but are limited because the need to obtain firmware for a particular device is hard. This project will address the challenge of *transfer fuzzing*. The idea is based on the observation that a device is expected to function similarly to its counterparts. As a result, if we can obtain firmware samples from the counterparts, we can fuzz test these samples and apply the fuzzing results to guide the testing of the target device. Solutions will benefit clients/customers who need to secure IoT devices for industrial and commercial usage.

Project Description

Emotional artificial intelligence (EAI) is effective to identify the negative mood and distraction level of drivers by observing drivers and monitoring their emotional and physical state. The monitoring of movement, facial expression, voice, gestures, and biofeedback; raises privacy and ethical concerns for drivers and passengers in vehicles, which have impeded the proliferation of EAI. This new project studies positive and negative impacts of EAI applications on vehicles and propose privacy principles and technical solutions to mitigate security, privacy, and ethical risks in automotive EAI systems for the proliferation of EAI-based ADAS. New societal and technical solutions will address key problems in transportation and be deployed to reduce collisions and improve road safety.

Project Description

In-vehicle infotainment (IVI) has evolved from CD players to heavily computerized interfaces vulnerable to attacks. Various car manufacturers have their own IVI systems, but the Android Automotive IVI system has gained popularity due to the widespread use of the Android operating system on mobile devices. Numerous studies conducted over the years have highlighted practically exploitable vulnerabilities in Android Automotive and consistently pointed out a widespread and systematic lack of security measures in IVI systems. This project will develop robust and practical defenses mechanisms for the security of Android-based IVI system during runtime.

Project Description

This project will bolster the security of Hardware Security Modules (HSMs) that manage digital keys and perform cryptographic operations. The strategy is to transition these devices to a quantum-resistant model to confront potential threats from quantum computing advancements. As financial institutions handle vast amounts of sensitive data, the project addresses the risk of quantum computing breaking the cryptographic systems currently used by HSMs, which could leave this sensitive data vulnerable. Implementing a quantum-resistant model will future-proof institutions against threats, safeguarding their financial and reputational standing, and fostering customer trust. This project will strengthen data security in financial institutions in our rapidly advancing quantum computing era.

Project Description

Digital identity systems play a crucial role in various sectors, including e-commerce, healthcare, finance and government services, so ensuring their security is paramount. This project studies the challenges and emerging threats in digital identity systems and propose effective solutions to improve their security. A formal framework is to be developed to allow the construction of secure digital identity systems based on a set of high-level requirements.

Project Description

This project will develop a Machine Learning (ML)-based platform that leverages threat information and adversarial tactics to enhance and expedite CTI procedures. Using deep learning, the platform analyzes diverse threat information and employs text-generation capabilities to create professional reports for both technical and executive audiences. The platform supports the CS team by offering dialogue-based assistance, actionable recommendations, and conversations, thereby improving the quality of CS services and reducing the time needed for cyber triage.

Project Description

Swarm technologies refer to the use of multiple coordinated and decentralized autonomous entities such as drones to complete a mission and/or achieve an objective. They work collectively to achieve a task that is otherwise hard to achieve with a single entity. Each entity in a swarm follows a specific algorithm/rule and communicates with other entities to exhibit a complex behavior to achieve a task. Drone-based swarm technologies are being used in security and surveillance, search and rescue, environmental monitoring and precision agriculture, infrastructure inspection, package delivery to list a few. Due to the nature of communication between drones in a swarm and the unpredictability of the swarm behavior, the threat landscape is different with major risks offering unique security issues when compared to a single drone.

Project Description

Developing, delivering and managing hands on training for cybersecurity curriculum is an ecosystem gap for colleges and universities. The current situation is ad hoc. For the majority of universities, labs are developed by a teaching assistant and require students to download virtual machines and run labs on their own laptops. This approach introduces technical challenges for students getting environments configured, challenges tracking results and challenges maintaining labs through staff turnover. An alternative approach is for institutions to provide a lab infrastructure for students. With a managed infrastructure that challenge shifts to managing the infrastructure and resourcing the development of labs.