Saudi Cultural Missions Theses & Dissertations
Permanent URI for this communityhttps://drepo.sdl.edu.sa/handle/20.500.14154/10
Browse
103 results
Search Results
Item Restricted Sensing, Scheduling, and Learning for Resource-Constrained Edge Systems(Saudi Digital Library, 2025) Bukhari, Abdulrahman Ismail Ibrahim; Kim, HyoseungRecent advances in Internet of Things (IoT) technologies have sparked significant interest in developing learning-based sensing applications on embedded edge devices. These efforts, however, are challenged by adapting to unforeseen conditions in open-world environments and by the practical limitations of low-cost sensors in the field. This dissertation presents the design, implementation, and evaluation of resource-constrained edge systems that address these challenges through time-series sensing, scheduling, and classification. First, we present OpenSense, an open-world time-series sensing framework for performing inference and incremental classification on an embedded edge device, eliminating reliance on powerful cloud servers. To create time for on-device updates without missing events and to reduce sensing and communication overhead, we introduce two dynamic sensor-scheduling techniques: (i) a class-level period assignment scheduler that selects an appropriate sensing period for each inferred class and (ii) a Q-learning–based scheduler that learns event patterns to choose the sensing interval at each classification moment. Experimental results show that OpenSense incrementally adapts to unforeseen conditions and schedules effectively on a resource-constrained device. Second, to bridge the gap between theoretical potential and field practice for low-cost sensors, we present a comprehensive evaluation of a sensing and classification system for early stress and disease detection in avocado plants. The greenhouse deployment spans 72 plants in four treatment categories over six months. For leaves, spectral reflectance coupled with multivariate analysis and permutation testing yields statistically significant results and reliable inference. For soils, we develop a two-level hierarchical classification approach tailored to treatment characteristics that achieves 75–86\% accuracy across avocado genotypes and outperforms conventional approaches by over 20\%. Embedded evaluations on Raspberry Pi and Jetson report end-to-end latency, computation, memory usage, and power consumption, demonstrating practical feasibility. In summary, the contributions are a generalized framework for dynamic, open-world learning on edge devices and an application-specific system for robust classification in noisy field deployments. These real-world deployments collectively outline a practical framework for designing intelligent, cloud-independent edge systems from sensing to inference.15 0Item Restricted Deep Learning based Cancer Classification and Segmentation in Medical Images(Saudi Digital Library, 2025) Alharbi, Afaf; Zhang, QianniCancer has significantly threatened human life and health for many years. In the clinic, medical images analysis is the golden stand for evaluating the prediction of patient prog- nosis and treatment outcome. Generally, manually labelling tumour regions in hundreds of medical images is time- consuming and expensive for pathologists, radiologists and CT scans experts. Recently, the advancements in hardware and computer vision have allowed deep-learning-based methods to become main stream to segment tumours automatically, significantly reducing the workload of healthcare professionals. However, there still remain many challenging tasks towards medical images such as auto- mated cancer categorisation, tumour area segmentation, and relying on large-scale labeled images. Therefore, this research studies theses challenges tasks in medical images proposing novel deep-learning paradigms that can support healthcare professionals in cancer diagnosis and treatment plans. Chapter 3 proposes automated tissue classification framework called Multiple Instance Learning (MIL) in whole slide histology images. To overcome the limitations of weak super- vision in tissue classification, we incorporate the attention mechanism into the MIL frame- work. This integration allows us to effectively address the challenges associated with the inadequate labeling of training data and improve the accuracy and reliability of the tissue classification process. Chapter 4 proposes a novel approach for histopathology image classification with MIL model that combines an adaptive attention mechanism into an end-to-end deep CNN as well as transfer learning pre-trained models (Trans-AMIL). Well-known Transfer Learning architectures of VGGNet [14], DenseNet [15] and ResNet[16] are leverage in our framework implementation. Experiment and deep analysis have been conducted on public histopathol- ogy breast cancer dataset. The results show that our Trans-AMIL proposed approach with VGG pre- trained model demonstrates excellent improvement over the state-of-the-art. Chapter 5 proposes a self-supervised learning for Magnetic resonance imaging (MRI) tu- mour segmentation. A self-supervised cancer segmentation framework is proposed to re- duce label dependency. An innovative Barlow-Twins technique scheme combined with swin transformer is developed to perform this self supervised method in MRI brain medical im- ages. Additionally, data augmentation are applied to improve the discriminability of tumour features. Experimental results show that the proposed method achieves better tumour seg- mentation performance than other popular self- supervised methods. Chapter 6 proposes an innovative Barlow Twins self supervised technique combined with Regularised variational auto-encoder for MRI tumour images as well as CT scans images segmentation task. A self-supervised cancer segmentation framework is proposed to reduce label dependency. An innovative Barlow-Twins technique scheme is developed to represent tumour features based on unlabeled images. Additionally, data augmentation are applied to improve the discriminability of tumour features. Experimental results show that the pro- posed method achieves better tumour segmentation performance than other existing state of the art methods. The thesis presents four approaches for classifying and segmenting cancer images from his- tology images, MRI images and CT scans images: unsupervised, and weakly supervised methods. This research effectively classifies histopathology images tumour regions based on histopathological annotations and well-designed modules. The research additionally comprehensively segments MRI and CT images. Our studies comprehensively demonstrate label-effective automatic on various types of medical image classification and segmentation. Experimental results prove that our works achieve state-of-the-art performances on both classification and segmentation tasks on real world datasets7 0Item Restricted EXPERIMENTAL STUDY OF THE IMPORTANCE OF DATA FOR MACHINE LEARNING-BASED BREAST CANCER OUTCOME PREDICTION(Saudi Digital Library, 2024) Yamani, Wid; Wojtusaik, JanuszEXPERIMENTAL STUDY OF THE IMPORTANCE OF DATA FOR MACHINE LEARNING-BASED BREAST CANCER OUTCOME PREDICTION Wid Yamani, Ph.D. George Mason University, 2025 Dissertation Director: Dr. Janusz Wojtusiak Researchers have used various large-scale datasets to develop and validate predictive models in breast cancer outcome prediction. However, a notable gap exists due to the lack of a systematic comparison among these datasets regarding predictive performance, feature availability, and suitability for different analytical objectives. While each dataset has unique strengths and limitations, no comprehensive studies evaluate how these differences impact model performance, particularly across diverse timeframes, survival, and recurrence outcomes. This gap limits researchers in making informed choices about the most appropriate dataset for specific research questions. Effective modeling and prediction of breast cancer outcomes (such as cancer survival and recurrence) rely on the dataset's quality, the pre-processing techniques used to clean and transform data, and the choice of predictive models. Therefore, selecting a suitable dataset and identifying relevant variables are as crucial as the choice of the model itself. This thesis addresses this gap by systematically comparing five prominent datasets for predicting breast cancer outcomes. This dissertation compares five datasets—SEER Research 8, SEER Research 17, SEER Research Plus, SEER-Medicare, and Medicare Claims data—focusing on breast cancer survival and recurrence. It evaluates the predictive performance of each dataset using supervised machine learning methods, including logistic regression, random forest, and gradient boosting. The models were tested on metrics such as AUC, accuracy, recall, and precision, with gradient boosting delivering the most accurate results. The findings indicate that SEER-Medicare, which integrates cancer registry data with three years of retrospective claims, outperformed the other datasets, achieving AUCs of 0.891 for 5-year survival and 0.942 for 10-year survival. This dataset's inclusion of comprehensive health information, including pre-existing conditions and other claims data, makes it particularly valuable for outcome prediction. However, a drawback of SEER-Medicare is that it primarily includes patients aged 65 and older, as it is based on Medicare data. This limitation reduces its suitability for predicting outcomes in younger breast cancer patients, a significant subgroup with distinct risk factors and treatment responses. SEER Research Plus ranked second, offering data on patient demographics, breast cancer characteristics, staging, outcomes, and treatment, with AUC values of 0.877, 0.901, and 0.937 for 5-year, 10-year, and 15-year survival, respectively. SEER Research 17 and SEER Research 8 include patient demographics, breast cancer characteristics, and staging information but lack treatment details. SEER Research 17, which covers a larger population with more variables, yielded AUC values of 0.870 for 5-year survival, 0.897 for 10-year survival, and 0.920 for 15-year survival. SEER Research 8, which covers a smaller population over a more extended period, yielded slightly lower AUC values of 0.857, 0.868, and 0.880 for 5-year, 10-year, and 15-year survival, respectively. Results indicate that including treatment and additional variables significantly enhances prediction accuracy while the data size is less critical. This thesis is the first study that compares SEER datasets and provides a groundbreaking, comprehensive evaluation of these datasets, providing crucial insights into how data characteristics influence breast cancer outcome modeling.15 0Item Restricted Stress Detection: Leveraging IoMT Data and Machine Learning for Enhanced Well-being(Saudi Digital Library, 2025) Alsharef, Moudy Sharaf; Alshareef, Moudywe focus on the detection of acute stress, characterized by short-term physiological changes such as changes in heart rate variability (HRV), breathing patterns, and other bodily functions. Often measurable through wearable or contactless sensors. Accurate detection of acute stress is crucial in high-pressure environments, such as clinical settings, to reduce cognitive overload, prevent burnout, and minimize errors. Current research on stress detection faces multiple challenges. First, most proposed methods are not designed to identify stress in unseen subjects, limiting their generalizability and practical applicability. Second, due to the sensitive nature of stress-related physiological data and the risk of data leakage, insufficient attention has been paid to ensuring data privacy while preserving utility. Third, many existing studies rely on synthetically induced stress in controlled environments, overlooking real-world scenarios where stress can have severe consequences. Finally, nearly all research in this domain employs invasive IoMT sensors or wearable devices, which may not be practical or scalable for real-world applications. This thesis presents five key contributions in the field of stress detection using Internet of Medical Things (IoMT) sensors and machine learning. First, it introduces a deep learning model based on self-attention (Transformer), trained and evaluated using the WESAD dataset, a widely used benchmark collected from 15 participants under controlled stress tasks. The model achieved 96% accuracy in detecting stress and was validated using leave-one-subject-out (LOSO) cross-validation to demonstrate generalizability to unseen individuals. Second, to ensure data privacy, a differential privacy framework was integrated into the model. This approach adds noise during training to prevent sensitive data leakage and achieved 93% accuracy, confirming it is both private and effective. Third, the thesis introduces a new dataset called PARFAIT, collected from 30 healthcare workers during real hospital duties (ICU, ER, OR) using non-invasive HRV sensors and the Maslach Burnout Inventory (MBI) to label stress levels. This dataset supports real-world analysis of stress among physicians. Fourth, a cost-sensitive model is developed using XGBoost and the PARFAIT dataset, assigning higher penalties to stress misclassifications that could lead to medical errors. This model achieved 98% accuracy and reduced false negatives, making it suitable for clinical settings. Finally, a contactless radar-based system is presented to detect stress using ultrawideband (UWB) radar, capturing HRV and breathing data. A deep learning model achieved 92.35% accuracy, offering a non-wearable, scalable alternative. Although the radar-based model achieved a slightly lower accuracy (92.35%) compared to the wearable-based model (96%), it provides several important advantages. It works with out any physical contact, helps maintain user privacy, and can be more practical to deploy in clinical settings where wearable sensors may not be suitable. The small drop in accuracy is mainly due to the limitations of radar in measuring HRV precisely. However, by combining radar-based HRV with breathing features, the overall performance remains competitive. 314 0Item Restricted The Role of Artificial Intelligence in Strengthening Cyber Defense Mechanisms: Opportunities and Challenges(University of Bedfordshire, 2024) Alanazi, Mohammed; Garner, LeeThis study explores the role of Artificial Intelligence (AI) in strengthening cyber defense mechanisms, focusing on the opportunities and challenges it presents. In recent years, AI has shown potential in enhancing threat detection, response efficiency, and proactive cybersecurity measures. The study examines various AI applications in cyber defense, including machine learning for real-time threat identification and natural language processing for analyzing large-scale data patterns. While AI provides significant advantages in mitigating cyber threats, challenges such as model interpretability, ethical concerns, and vulnerability to adversarial attacks persist. The findings contribute to cybersecurity by highlighting both the promising capabilities and limitations of AI in this domain, suggesting future research directions to address these challenges.21 0Item Restricted Cloud-Computing for Carbon Finance Decision Support System with Dynamic Machine Learning Repository Management Service(Saudi Digital Library, 2025-06-20) ALTHUNAYYAN, AZZAM; Dong, Yuan; Liu, Charles ZThis thesis presents the development of a local machine learning-based decision support system designed to predict future carbon prices. Carbon finance markets play a critical role in supporting global climate change mitigation strategies, where market price volatility poses substantial challenges for investors, policymakers, and carbon traders. This research integrates and compares five advanced forecasting models — Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Categorical Boosting (CatBoost), Facebook Prophet, and Autoregressive Integrated Moving Average (ARIMA)—by training and evaluating them on historical carbon price datasets to identify the optimal predictive approach. The system is implemented as a Spring Boot web application operating in a local environment, serving as a functional proof-of-concept for potential future deployment on scalable cloud infrastructure. The models are evaluated using performance metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and the coefficient of determination (R²). Results demonstrate that the selected best-performing model offers superior forecasting accuracy and robustness under varying market conditions. This work contributes to the intersection of carbon finance and artificial intelligence by delivering an extensible, locally operable system that lays the groundwork for future cloud-based deployment, supporting informed decision-making for stakeholders in the carbon trading ecosystem.35 0Item Restricted Characterisation of Nanoclusters on Surfaces Using Scanning Transmission Electron Microscope and Machine Learning(University of Birmingham, 2025) Alhabeadi, Hanan Hamed; Theis, WolfgangSingle metal atoms (SMAs) and metal nanoclusters (MNCs) have attracted considerable interest in recent years due to their unique properties; their properties at the surface are important in many potential applications, such as catalysis, sensing, and thin film fabrication. However, it remains unclear what determines metal cluster formation in this process. Therefore, a series of experiments have been designed and performed to address the mystery of filtered and non-filtered metal clusters formed in magnetron sputtering with a size-selected cluster source using several metals, tantalum (Ta), silver (Ag), copper (Cu), and platinum (Pt). We describe how techniques for answering these questions are developed by focusing on mechanisms governed by gas pressure, magnetron power, temperature, substrate voltage, condensation length, and cooling the system. This study employs high-resolution aberration-corrected scanning transmission electron microscopy (AC-STEM) to acquire images with a sufficient spatial resolution to separate individual atoms and characterise metal clusters deposited on different supports. It is possible to study atoms/clusters on surfaces directly by capturing annular dark field (ADF)-STEM images at atomic resolution. This thesis focuses on studying the distribution of single atoms and nanoclusters on the surface under different conditions. Furthermore, STEM stereo imaging was employed to examine the height of Ta nanoclusters on different substrates. Additionally, two machine learning models were developed and evaluated in order to predict the position of clusters in a single simulated image and the parallax shifts in stereo images.21 0Item Restricted Cross Dataset Fairness Evaluation of Transformer Based Sentiment Models(Saudi Digital Library, 2025-05-10) Zuiran, Sara; Bhattacharyya, SiddharthaWith the growing exploration of Natural Language Processing (NLP) systems in decision-making environments, it is essential to evaluate technical and ethical aspects of the dataset and the NLP model to improve fairness. To assess fairness, the thesis examines demographic imbalances in sentiment classification models by evaluating transformer-based models fine-tuned on the Stanford Sentiment Treebank version 2 dataset (SST-2) against the demographically annotated Comprehensive Assessment of Language Model dataset (CALM). This work identifies performance disparities in sentiment prediction across demographic groups by examining sensitive attributes such as gender and race. The study evaluates both the RoBERTa and MentalBERT transformer models using a complete set of fairness metrics consisting of Statistical Parity Difference (SPD), Equal Opportunity Difference (EOD), False Positive Rates (FPR), False Negative Rates (FNR), Jensen-Shannon Divergence (JSD), and Wasserstein Distance (WD). The analysis examines both group-vs-rest and pairwise subgroup comparisons, including gender and ethnicity. Results show that applying adversarial mitigation reduced fairness disparities across demographic subgroups, with the most notable improvements observed for non-binary and Asian users. The observed disparities emphasize the challenge of reducing performance gaps across demographic subgroups in sentiment classification tasks. The thesis introduces a practical framework for evaluating demographic dis- disparities, extends fairness analysis, and assesses the impact of mitigation techniques in cross-dataset sentiment classification. This research proposes a framework that demonstrates a path toward creating inclusive NLP systems and establishes the groundwork for upcoming ethical Artificial Intelligence (AI) studies.13 0Item Restricted Novel Framework for Integrating Blockchain Technology into Logistics and Supply Chain Services(Saudi Digital Library, 2025) Alkhaldi, Bidah; Al-Omary, AlauddinBottlenecks and operational inefficiencies in supply chains still persist despite technological innovations, due to structural and managerial issues. Blockchain integration presents a viable solution to these long-standing issues by offering tamper resistant ledgers, secure transactions and automation capabilities. This research takes a novel approach by designing a blockchain integration framework for supply chains, modifying the MOHBSChain framework to create the Supply-Blockchain framework. This framework is validated by developing a functional prototype using Hyperledger Fabric, by considering a port decongestion use case scenario. This research adopted an inductive approach, starting with informal observations of real-world port operations and a targeted literature review to identify patterns and challenges. The framework development was guided by the principles of transaction cost economics, resource-based view, and diffusion of innovations theories. MoSCoW method was used to prioritize features, while agile project management was adopted to ensure timely completion. Hyperledger Firefly and its connector framework were used as the middleware to facilitate blockchain integration, while chaincode developed using Go language was packaged and deployed to implement smart contracts. Raft orderer consensus mechanism was chosen to ensure resilience and fault tolerance. From a core functionality standpoint, the prototype allows initiation of smart contracts corresponding to functions such as creating and editing supply chain process-related documents, minimizing manual interventions and enhancing efficiency to reduce port congestion. It also offers live tracking of blockchain transactions, facilitating transparency and oversight, the permissioned nature of Hyperledger Fabric ensures security and robust access controls. Results of functional and performance testing conducted using Hyperledger Caliper, Prometheus, and Grafana, were satisfactory; this indicates the prototype's potential in alleviating bottlenecks in supply chains and quickly delivering benefits to key stakeholders such as port authorities, customs officers, shipping line representatives and logistics providers. In terms of limitations, the prototype is limited to basic functionalities and lacks advanced features required to meet operational and regulatory standards. Future improvements can focus on integrating AI for tasks such as predictive analytics and automated document verification, while technologies such as NFT-based schemas can enhance ownership verification and improve asset tracking.39 0Item Restricted Investigating Factors That Affect Student Engagement and Academic Performance in Novice Students in CS Education(Saudi Digital Library, 2025-05-30) Albakri, Sultanah Abdullah A; Ada, Mireilla BikangaIn recent years, improving student engagement in Computer Science Education (CSE) has gained significant attention. Despite the abundance of research on student engagement, the relationships between the dimensions of student engagement (behavioural, cognitive, emotional, and social), students’ confidence in learning Computer Science (CS), and their beliefs in the usefulness of learning CS have not been thoroughly investigated in CSE. Thus, the primary objective of this thesis is to address the above arguments by examining multidimensional student engagement factors and identifying factors influencing engagement and CS learning performance among novice students in CS courses. This thesis uses a mixed methods approach and includes six research studies building on each other. The research was conducted between 2021 and 2024. It was carried out in various educational institutions, including schools and universities, across two countries: Saudi Arabia and Scotland. This thesis is structured into five phases, starting with a systematic literature review (SLR) to investigate previous research on student engagement in CSE at both school and higher education (HE) levels. The review aimed to identify study objectives, methods, and different indicators used to measure student engagement in CSE. This SLR led to the second phase, which includes two studies focusing on confidence and perceived usefulness factors. The first study explored the relationship between different dimensions of student engagement and the confidence and perceived usefulness among female high school students in Saudi Arabia. The second study applied ML algorithms to analyse engagement indicators and predict these two factors among female students. The third phase involves a quantitative study using structural equation modeling to examine the relationships between CS academic performance, student engagement, confidence, and perceived usefulness. The study also explores and compares student engagement levels between two different groups of students, considering academic cultural contexts and gender differences. The fourth phase includes a qualitative study examining novice students’ perspectives on factors that affect their engagement and academic performance in CS classes. The last phase includes a study that explored student engagement, log data (learning analytics) and used ML prediction models to explore and determine the predictive power of different engagement indicators that could be used to predict CS learning performance. The findings support the thesis claim that student engagement, which comprises behavioural, cognitive, emotional, and social dimensions, and is influenced by students’ confidence in learning CS and their perceived usefulness of learning CS, affects CS learning performance. This thesis makes several important contributions to the CS field. A key contribution is the development of confidence, perceived usefulness, and student engagement model to enhance CS learning performance (CUSEL). This model could be used by CS educators to support the learning and teaching process in CS courses. The thesis provides implications for educators, researchers, and any stakeholders who want to improve or design effective interventions that would increase engagement to improve student learning outcomes in CSE.17 0