https://www.um.edu.mt/library/oar/handle/123456789/141711 Wed, 03 Jun 2026 20:19:19 GMT 2026-06-03T20:19:19Z https://www.um.edu.mt/library/oar/handle/123456789/146924 Title: Advancing automated Maltese spell checking using deep learning Abstract: Research in Natural Language Processing (NLP) for Maltese has advanced in recent years, yet reliable and contextually aware spell checking support remains limited. Existing tools, such as spelling.mt, provide dictionary-based correction but cannot handle grammatical or context-sensitive errors, leaving a clear gap com pared with high-resource languages. This dissertation investigates whether a Large Language Model (LLM) can be fine-tuned to develop an effective and resource efficient spell checker for Maltese. The study adapts Meta’s Large Language Model MetaAI (LLaMA)-3-8B-Instruct model using pairs of incorrect and corrected Maltese words and sentences. A cus tom fine-tuning corpus was created by introducing linguistically motivated syn thetic errors into correctly written text and incorporating smaller sets of authen tic incorrect-correct pairs from prior research. Two complementary models were trained: a word-level model that corrects individual tokens and a sentence-level modelthatcorrectsgrammaticalandorthographicerrorswithinfullsentences. Their performance wasevaluated across synthetic, real-world, and correct-input datasets. The main contribution of this dissertation is the development of the first LLM based spell checking models, specifically fine-tuned for Maltese, supported by a curated error-correction dataset and a systematic multi-setting evaluation. To our knowledge, this study is the first to explore the use of a modern generative LLM to build a Maltese spell checker and assess its potential in a low-resource setting. The developed models demonstrated strong performance on incorrect-correct pairs with synthetically generated errors, with the word-level model achieving 84.64% accuracy and the sentence-level model reaching 95.3%. For real-world eval uation, only the sentence-level model was tested because no suitable word-level datasets are available. In this context, the accuracy of the sentence-level model dropped to 29.6%, reflecting the increased variability of naturally occurring mis takes. However, metrics such as Error Annotation Toolkit (ERRANT), Bilingual Evaluation Understudy (BLEU), and Grammar Language Evaluation Understudy (GLEU) returned promising values, indicating that the model often produced cor rections close to the target. Overall, the findings demonstrate that fine-tuning an LLM offers a promising pathway towards effective Maltese spell checking, while revealing key challenges related to generalisation and data scarcity. The study provides an initial bench mark andoutlines concrete directions for developing more robust, deployable spell checking solutions for low-resource languages. Description: M.Sc.(Melit.) Thu, 01 Jan 2026 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/146924 2026-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/146923 Title: Enhancing transparency and interpretability in AI-driven algorithmic trading Abstract: Algorithmic trading increasingly relies on AI-driven decision systems, yet opaque models limit trust and accountability. This study investigates how Reinforcement Learning (RL) can be made more transparent by combining a model-agnostic explainability framework with Large Language Model (LLM) based narrative synthesis. The framework comprises four layers linking trading behaviour to feature attribution and temporal dynamics, stability and regime sensitivity, policy surrogacy, and reward decomposition. In Experiment 1, we apply state-of-the-art RL algorithms to constituents of the Dow Jones Industrial Average and evaluate performance using standard return and risk measures. We investigate modern explainability techniques across market regimes to characterise which indicators drive decisions, how stable attributions are over time, and how policies can be approximated by compact surrogate rules. The results indicate convergent feature drivers, expected masking behaviour, smoother attributions, indicating that explanations are temporally stable, and low-complexity surrogates with credible fidelity. In Experiment 2 we extend these findings by investigating how explainability artefacts can be translated into grounded natural-language narratives. Explanations generated from the framework are assessed through automated text metrics and a human-centred study with participants at varying levels of trading experience. Our results show that structured prompting improves lexical quality and adherence to factuality constraints relative to a zero-shot baseline. An important finding is that participants view the combined visual–narrative explanations as clear, moderately trust worthy, and practically helpful, with narratives particularly useful for reconstructing the agents’ reasoning and feature contributions. The contributions of this work are threefold. First, the study introduces a unified, model-agnostic explainability framework for financial RL linking feature-level attributions to policy behaviour and realised rewards. Second, it proposes and validates a pipeline for grounded natural-language synthesis anchored in quantitative explainability outputs. Third, in daily DJIA trading and a small human study, it provides evidence that technical faithfulness and human-centred accessibility can be advanced together, offering a template for transparent decision support in finance. Description: M.Sc.(Melit.) Thu, 01 Jan 2026 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/146923 2026-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/146919 Title: Towards optimal project planning : a unified framework for story point and effort estimation in agile software development Abstract: The persistent challenge of accurate story point and effort estimation in Agile Software Development stems from the inherent subjectivity and variability of conventional approaches such as Planning Poker and expert judgement. This investigation tackles these fundamental limitations through the development of a hybrid deep learning architecture that combines LSTM networks, CNN layers, and Transformer encoder blocks, aiming to support a more systematic, data-driven estimation methodology in place of purely subjective, human-driven approaches. The proposed approach leverages multi-task learning paradigms to concurrently predict both story points and effort estimates, capitalizing on their intrinsic interdependencies to enhance predictive accuracy. The framework implements systematic data preprocessing pipelines and employs a concatenation-based combination of GloVe and Word2Vec embeddings to capture both global and local semantic relationships, while integrated attention mechanisms support transparency and interpretability in estimation decision-making processes. Comprehensive evaluation is conducted using the TAWOS dataset, encompassing 458,232 issues across 39 diverse open-source projects. The LSTM-CNN-Transformer methodology is rigorously compared against established traditional techniques and previously proposed AI approaches from the literature under a unified experimental protocol, employing a consistent suite of evaluation metrics, including MAE, MedAE, accuracy, F1-scores, and MMRE. Across the majority of projects, the proposed model matches or outperforms competing methods for both story point classification and effort prediction, demonstrating robust and generalisable performance rather than isolated improvements on a few favourable cases. This research makes several contributions to both theoretical understanding and practical implementation in software estimation, delivering a comprehensive data-driven solution that reduces subjectivity whilst preserving interpretability and ensuring practical applicability for Agile development teams. Description: M.Sc.(Melit.) Thu, 01 Jan 2026 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/146919 2026-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/146915 Title: AI in large search spaces with a domain agnostic approach Abstract: This dissertation presents a systematic comparison of classical and quantum machine learning for solving the 3×3×3 Rubik’sCube, a benchmark for high-dimensional discrete search spaces. The classical approach uses an optimised QUBE reinforcement learning architecture with structured state representations, distance-based reward shaping, and Bayesian hyperparameter optimisation. The quantum approach employs a 6-qubit Variational Quantum Circuit (VQC) with amplitude encoding trained end-to-end on full-cube states. A hybrid quantum classical architecture is also evaluated. Results show the classical system achieves 65% System Success Rate (SSR) with stable solution trajectories. The quantum VQC demonstrates functional learning and extreme parameter compactness but achieves only 12% SSR due to circuit-depth limitations, barren plateau effects, and computational costs. The hybrid architecture achieves 15% SSR. The findings provide rigorous comparison under identical evaluation conditions, highlighting optimised classical RL’s strong performance, NISQ era quantum constraints, and hybrid architectures’ potential as hardware and algorithms mature. Description: M.Sc.(Melit.) Thu, 01 Jan 2026 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/146915 2026-01-01T00:00:00Z