http://hdl.handle.net/1893/36 Electronic theses of Computing Science and Mathematics students. Mon, 28 Sep 2026 10:20:36 GMT 2026-09-28T10:20:36Z http://hdl.handle.net/1893/37730 Title: Understanding Malware using Ontology-based knowledge graph and deep learning techniques Author(s): Roy Chowdhury, Ipshita Abstract: Due to rapid growth and advancement of Big Data and Internet of Things (IoT), industries and organizations are being the main targets of cyber criminals. As IoT enables the linking of all types of `smart' devices industrial appliances, personal information or sensitive data are being constantly attacked by novel malware variants. Therefore, in recent years cyber security has become a major concern due to the emergence of new advanced malware mutants which are causing harm to the users for financial gain to personal information loss. Due to transportation and sharing of huge amount of data, internet-enabled devices and social media platforms are becoming more vulnerable to new generation malware attacks. Because of the limitations of traditional Machine Learning based approaches, it's very difficult to deal with the new malware generations with complex behaviour and varying code structures. Different categories of malware families represents complex, dynamic behaviours and characteristics which can cause a novel and targeted attack in a cyber-system. Existence of large volume of malware types with frequent new additions hinders the cyber resilience effort. This motivation leads to develop a new ontology driven framework that can capture recent malware behaviours and construct novel malware detection and classification framework for new generation malware. This research built an ontology-based knowledge graph to capture metamorphic malware behaviour in the form of API call sequences, and it constructed a dynamic model for implementing an API-based knowledge graph from the domain ontology. The ontological structure are used as the domain knowledge to develop several deep learning algorithms have been incorporated for the prediction of new generation malware classes. In this research a number of deep learning techniques were developed for malware analysis by incorporating CNNs, ResNet-50, GANs, achieving higher accuracies than state-of-the-art models. CNNs and other deep learning architectures were implemented as standalone models and by integrating with other models as well. To address the limitations of static deep learning models, this work includes continual learning approaches to avoid catastrophic forgetting while promoting adaptability in dynamic threat environments. It further addresses two new approaches bringing ontology to Deep Learning: Domain Adaptation, transferring knowledge from one source to one target domain with results pooled in the subsequent stage and then Building Knowledge Graphs using Deep Learning-based Image Classification and Ontology Integration to better support semantic representation through structured integration of the output of classifications into structured knowledge. The overall research work has been carried out to built a meaningful domain specific ontological framework, a wide range of Deep Learning-based neural architectures have been implemented on publicly available benchmark datasets. Employment of various advanced deep learning techniques to improve accuracy and address key challenges such as catastrophic forgetting. Integration of deep learning and ontology with Domain Adaptation and construction of three deep learning based malware knowledge graphs for enhancing interpretability and advancing AI-driven malware analysis in cybersecurity. Thu, 27 Mar 2025 00:00:00 GMT http://hdl.handle.net/1893/37730 2025-03-27T00:00:00Z http://hdl.handle.net/1893/37138 Title: Beyond Intuition: Redefining Internal Audit Risk Assessments Through Machine Learning Paradigms Author(s): Shivram, Vivek Abstract: Established literature has found that Internal Audit Functions (IAF) increasingly view AI- powered innovation as being indispensable to continue staying relevant in the light of in- creased organisational complexity and regulatory expectations. This thesis explores opportunities to deploy Machine Learning (ML) capabilities to assess how IAFs can benefit from increased precision, objectivity and efficiency for delivering internal audit risk assessments. Specifically, this thesis explores how the risk assessment process, a fundamental component of the audit lifecycle, can be redefined through ML to generate measurable value for the IAF. In doing so, this thesis makes several researched assumptions and acknowledges the impact of relevant limitations. Three research questions were framed for evaluation throughout the thesis, to explore the relevance and findings of prevailing literature, identify suitable ML techniques for practical implementation, and measure value from any proposed implementations. In doing so, the thesis supplements the findings of prior academic research, whilst laying the groundwork for further practical improvement by other IAFs. The thesis applied multiple ML techniques to answer these research questions, combining academic ML principles with real-world datasets and implementation techniques, offering a meaningful starting point for IAFs looking to improve their approach to internal audit risk assessments. Three rounds of experiments were performed to reframe risk assessments as specific problem statements for ML models, supported by (1) an examination of the ex- tant risk assessment methodology used by the Senior Leadership Team (SLT) at the subject organisation, a world-leading financial services firm, (2) an analysis of organisational data relevant to the risk assessment, anonymised for the purposes of this thesis, and (3) the construction and application of several ML hypotheses on this dataset. The conclusions of these experiments were consolidated to answer the research questions, including an assessment of how the value arising from such efforts could be measured for continued sponsorship and improvement. Overall, the thesis noted that there is no single best ML approach for practical implementation in all cases, and that the choice of ML models would vary based on the quality of underlying data, consistent audit methodologies, as well as the size of the dataset including the final set of input features. The thesis found that a supervised machine learning approach was most suitable for internal audit risk assessments, given the high class-specific precision ratio of the final model (c.88%) on the test dataset. However, the thesis noted significant dependencies on high-quality datasets and a requirement to carefully curate relevant features to ensure meaningful model outputs. Although the thesis evaluated unsupervised machine learning techniques, these were discounted from practical implementation due to comparatively lower accuracy, inhibiting practical implementation for the given dataset. However, the thesis also noted that such techniques may generate better results in alternative practical settings, where significantly larger datasets are used for risk assessment cycles. Wed, 01 Jan 2025 00:00:00 GMT http://hdl.handle.net/1893/37138 2025-01-01T00:00:00Z http://hdl.handle.net/1893/37134 Title: Machine Learning-Driven Sentiment Analysis of UK CBDC Tweets Versus Thematic Evaluation of Public and Institutional Perspectives: A Comprehensive Study on the Digital Pound Author(s): Kaur, Guneet Abstract: In an era of rapid digital transformation and declining cash usage, central banks worldwide are exploring digital currencies to modernise monetary systems and maintain monetary sovereignty. The proposed digital pound in the UK has ignited intense debate among policymakers and citizens, particularly concerning issues of privacy, surveillance, and economic inclusion. Against this backdrop, this study investigates how public sentiment evolves in response to policy milestones and how it aligns — or diverges — from official Bank of England (BoE) narratives concerning a potential digital pound, addressing a critical knowledge gap in understanding public reception of central bank digital currencies (CBDCs). To meet this objective, the study adopts a novel interdisciplinary approach integrating advanced sentiment analysis using fine-tuned transformer models (DistilBERT, RoBERTa, XLM-RoBERTa), communication theories (e.g., framing theory, agenda-setting theory, and Grunig’s two-way symmetrical model), and analysis of policy messaging. A bespoke, domain-specific gold-standard dataset was created and validated, enabling the fine-tuning of these models. RoBERTa, trained for three epochs, emerged as the optimal model for classifying nuanced discussions related to the digital pound. Longitudinal analysis of public discourse on X (formerly Twitter) across three key periods (2020, 2023, and 2024), corresponding to major BoE policy announcements, revealed an “Exploration–Polarisation– Adaptation” sequence: initial cautious optimism evolved into pronounced negativity, particularly concerning privacy and government control, following major policy announcements, with a partial rebound after official BoE responses. A comparative thematic analysis with official BoE narratives highlighted key discrepancies, notably a “privacy framing gap” where the BoE's technically focused approach to data protection diverged from public concerns over surveillance and government overreach. This mismatch underscores a disconnect between the technocratic framing of policy narratives and public anxieties, pointing to the imperative for two-way symmetrical communication to establish trust. By illustrating how official narratives can both shape and overlook public views, this study contributes practical insights for policymakers and researchers navigating the complex interplay of technology, policy communication, and public opinion surrounding the digital pound. Recommendations include targeted public engagement on privacy, transparent implementation roadmaps, and a shift towards two-way symmetrical dialogue. While acknowledging limitations related to data source representativeness and the potential influence of external factors, this study provides a comprehensive, empirically grounded understanding of public sentiment dynamics in digital monetary policy. Future research should extend these findings by incorporating broader data modalities and demographic insights, and by applying automated hyperparameter optimisation techniques to further refine understanding of public sentiment dynamics around the digital pound and similar CBDC innovations. Sat, 01 Feb 2025 00:00:00 GMT http://hdl.handle.net/1893/37134 2025-02-01T00:00:00Z http://hdl.handle.net/1893/37074 Title: Improving the Scalability of 6TiSCH Wireless Industrial Networks using Scalable Scheduling Reservation Protocol Author(s): Kumar, Kaushal Abstract: The Internet of Things is currently evolving. The demand for scalable and high throughput low-power wireless sensor networks is transforming the concept of automation of industrial processes. IPv6 is considered a potential solution allowing a large number of sensor devices connected to multi-hop low power wireless sensor networks to exchange information over a range of 1-2 km and without being dependent on infrastructure. TSCH over IEEE 802.15.4 standard is a proposal at the MAC layer in the low power IPv6 protocol suite called 6TiSCH. However, the implementation of TSCH has been subject to poor scalability in the 6TiSCH networks beyond 50 nodes where nodes frequently appear and disappear. The main cause is poor scheduling of link-layer resources. In 6TiSCH networks, the reservation of TSCH cells without a context leads to under-or-over estimation of actual bandwidth requirements, while monitoring the buffer condition for traffic adaption leads to a high level of additional overheads. Existing proposals employ an ‘On the fly’ reservation approach using a fixed threshold to tune up cell consumption and incur performance trade-offs. More generally, cell selection in TSCH-led scheduling has been a key challenge, as allocating a Tx cell closer to each other in TSCH slotframe reduces delay; however, this causes collisions during transmission to increase. Furthermore, delay is also increased by not allocating a sufficient volume of cells to a node probing the shortest path to the root. Existing approaches adapt cell selection based on the requirements of the application. Apart from poor traffic adaption, and inefficient cell selection, a fixed distribution of traffic in the network undermines the ability of nodes to adapt their behavior according to demand as some nodes may be able to transmit higher payload than the others depending on their distance to the root and volume of overprovisioned cells. This inability negatively affects propagation. Existing algorithms have not addressed this issue. This thesis introduces the Scalable Scheduling Reservation Protocol (SSR) to tackle these issues. SSR uses an analytical technique called cake-slicing for traffic adaptation, prioritizing higher resource allocation to nodes closer to the network root. It employs schedule compactness via cell selection and collision-free scheduling for faster, more reliable delivery, and integrates dynamic queue optimization to minimize congestion. While SSR delivers a strong proposal for medium-sized networks using lower consumption of TSCH resources, its performance in terms of reliability, a ratio between the number of data packets successfully received over the volume sent by a transmitter, declines in larger networks due to limited proliferation of information, causing fewer routes in the network. SSR is then implemented under hybrid scheduling design, aiming to further improve reliability. The results showed improved performance in terms of reliability within the network size of 100, compared to the minimal scheduling function, for all tested conditions. However, with more challenging traffic conditions, the performance still deteriorates. The main reason is the poor proliferation of information. To further improve the scalability, an increased penetration to shared resources (cells) is necessary. That is, any node under the common ancestry (in the topology) can negotiate for available cells. This functionality of distributed scheduling is incorporated in the proposed Decentralized and Broadcast-based Scalable Scheduling Reservation protocol (DeSSR), which exhibits high reliability under heavy traffic conditions, incurs low latency, and low consumption of TSCH cells compared to other solutions under this category. Mon, 22 Jan 2024 00:00:00 GMT http://hdl.handle.net/1893/37074 2024-01-22T00:00:00Z