Browsing Computing by Department
Now showing items 1-3 of 3
Crowdsourced linked data question answering with AQUACOLDThere is a need for Question Answering (QA) to return accurate answers to complex natural language questions over Linked Data, improving the accessibility of Linked Data (LD) search by abstracting the complexity of SPARQL whilst retaining its expressiveness. This work presents AQUACOLD, a LD QA system which harnesses the power of crowdsourcing to meet this need.
Quantisation feasibility and performance of RSS-based secret key extraction in VANETsVehicular Ad Hoc Networks (VANETs) has emerged as a unique implementation of Mobile Ad Hoc Networks (MANETs). These networks promise to increase road safety and improve the driving experience by exploiting recent advances in wireless technologies for both intra-vehicle and inter-vehicle communications. Physical layer security is a promising alternative approach to secure communication in VANETs where physical and applications' constraints encourage the use of lightweight and fast cryptographic algorithms. Our work focuses on the quantisation stage of the secret generation process, by reviewing existing schemes in the public domain and associated performance metrics. Evaluations are done through simulation with the aid of a wireless channel model which includes three-dimensional scattering and scatterers' mobility. Preliminary findings show that RSS-based algorithms do not perform efficiently in the proposed vehicular stochastic wireless model. Hence they are not able to satisfy the typical low latency required in safety-related broadcasting messaging. We conclude that more research is desirable to design protocols capable of taking advantage from the nodes' high-mobility and the consequent variability of both coherence intervals and level crossing rates, to further improve secret bit extraction throughput.
Understanding the cyber-victimisation of people with long term conditions and the need for collaborative forensics-enabled disease management programmesResearch shows that people with long term conditions and disabilities are frequently labelled as vulnerable, and commonly victimised online. They require instrumental support to understand their conditions and empower them to manage their own treatment in everyday life. However, additional short and long term consequences related to cyber-victimisation could intensify existing psychological and health complications. For instance, ‘distress’ as a commonly reported impact of cyber-victimisation could theoretically lead to neurohormonal changes in the blood, increasing cortisol, catecholamine and insulin secretion resulting in increased blood glucose, heartbeat, blood pressure, urination and other changes. Therefore, in this study we demonstrate the need and explain the means towards extending support and risk assessment systems and procedures to cover the collection and preservation of incidents reported by individuals. This can be used to support third-party interventions such as taking a legal action in cases where the impact of cyber-victimisation is seen to escalate and worsen. As such, we first define vulnerable groups with long term conditions and provide a review of the impact of various types of cyber-victimisation on their health management. Then, we discuss how Disease Management Programmes (DMP) developed over time to include web-based applications as an example of existing cost-effective approaches to improve the quality of healthcare provided to people with long term conditions. We then demonstrate the added value of incorporating forensics readiness to enable Police intervention, support the victim’s eligibility for extended instrumental support from national health services. Finally, this level of documentation offers an opportunity to implement more accurate methods to assess risk associated with victimisation.