https://hdl.handle.net/2123/345 Thu, 09 Jul 2026 17:06:24 GMT 2026-07-09T17:06:24Z https://hdl.handle.net/2123/35552 On Vassiliev Invariants and Weight Systems of Classical and Welded Knots Lin, Damian Vassiliev invariants are a special class of knot invariant that are analogous to polynomial functions on, for example, the real line. We review how Vassiliev invariants can be constructed from Lie algebra objects in arbitrary monoidal categories and are completely determined by their weight systems. We make some computations of the values of the g2 and f4 exceptional Lie algebra weight systems on a certain special class of chord diagrams. We generalise a construction of Hinich-Vaintrob to welded knots, thereby constructing a universal welded weight system. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35552 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35550 Efficient Edge-AI: Towards the Future of Implantable and Smart Medical Devices Herbozo Contreras, Luis Fernando Epilepsy affects over 1% of the global population and imposes substantial clinical, social, and economic burdens. Although pharmacological therapy is the first line of treatment, approximately 30–40% of patients remain drug-resistant, making neuromodulation one of the few viable options. However, current neuromodulation systems are largely open-loop or depend on cloud-based AI, limited by latency, power, privacy, and scalability. These constraints hinder autonomous, personalised, implantable closed-loop neurostimulation. This thesis investigates neuromorphic computing as a paradigm for next-generation closed-loop neuromodulation, focusing on seizure detection and prediction in epilepsy. It introduces neuromorphic neuromodulation and shows how biologically inspired, on-device intelligence can support self-responsive and personalised therapies. Building on this framework, the thesis develops seizure detection systems using liquid-time constant neurons and dendritic spiking mechanisms with heterogeneous temporal dynamics. These models enable efficient neural-signal representation without expensive feature extraction and show robust out-of-sample generalisation across patients and recording conditions on large-scale clinical EEG datasets. The thesis also addresses edge learning by proposing a neuromorphic learning rule for on-device adaptation toward patient-specific treatment under strict power and memory constraints. This mechanism can be adopted by the developed models to support personalised, low-latency intelligence at the edge. Finally, learnable activation functions are explored within artificial neural networks to improve training efficiency and interpretability, highlighting a pathway for integration with neuromorphic technology. Together, these contributions advance neuromorphic neurotechnology toward autonomous, personalised, and continuously learning closed-loop systems, with implications beyond epilepsy to a broader range of neurological disorders. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35550 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35549 Blockchain-based Advanced Information Infrastructure and its Applications in Demand-Side Energy Systems Yu, Teng The energy sector is moving from centralised one-way power systems to decentralised bidirectional smart grids driven by distributed energy resources (DERs). This shift requires infrastructure for secure coordination, verifiable markets, and prosumer data privacy. Existing blockchain systems have limited scalability, security, and interoperability, restricting high-frequency smart-grid applications. This thesis designs blockchain infrastructure for peer-to-peer (P2P) energy trading and power load forecasting (PLF). It has three theoretical innovation points. First, a dual-blockchain architecture with an Improved Optimistic Rollup (IOR) improves vertical throughput by offloading heavy computation from a primary to a secondary blockchain. Second, a Transaction Batch Generation (TBG) protocol for leaderless Byzantine Fault Tolerance (LBFT) consensus improves horizontal throughput by letting every node broadcast blocks, avoiding redundant transaction rebroadcast, and reducing transaction censorship to nearly zero. Third, a Blockchain-of-Blockchains (BoB) architecture, with Cross-Chain Token Exchange (CCTE) and Cross-Chain Data Interoperability (CCDI), supports asset and data transfer across blockchains with lower latency and memory overhead. The infrastructure is applied to two demand-side problems. For P2P market clearing, it is combined with Trusted Execution Environments (TEEs), D-TASK, and TEAR-DO to address the privacy-robustness-latency trilemma: prosumer data remain private, distributed optimisation preserves optimal convergence, and clearing time is reduced by an order of magnitude over state-of-the-art methods. For PLF, CTP-FL combines the infrastructure with message coding and commitment-based dual consensus to preserve local-model privacy, tolerate Byzantine faults at client and server sides, and keep communication latency below local model training time. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35549 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35548 Towards Geometry-Grounded World Understanding Tang, Liyao Understanding the three-dimensional (3D) world underpins embodied artificial intelligence, enabling robotics, autonomous driving, augmented and virtual reality, and, more broadly, spatial intelligence. Yet the most direct representation of a 3D scene can be deceptively simple: a point cloud, \ie, a set of Cartesian coordinates sampled from scene surfaces, from which objects and semantic structure must be inferred. In practice, point clouds are noisy, incomplete, and irregularly sampled, making reliable interpretation fundamentally challenging. Geometry-grounded world understanding demands semantics that are consistent with metric 3D geometry. Semantic segmentation provides a principled route from geometric measurements to high-level scene understanding. In 3D point clouds, semantic segmentation anchors geometry-grounded world understanding by coupling fine-grained semantics with explicit 3D geometry and confronting a core perceptual challenge: forming structured and coherent interpretation from noisy, incomplete, and irregular samples. This thesis investigates how explicit 3D geometry can be elevated from a passive input to a source of structure for learning and generalization. We treat geometry as a prior that constrains what a plausible segmentation should look like, modulates how noisy supervision should be used, and guides how models adapt when spatial statistics shift. Building on this view, we advance geometry-grounded scene segmentation along three complementary aspects of the learning problem: the output, the supervision, and the adaptation Overall, this thesis approaches scene segmentation by structuring the outcomes of learning, the supervision that guides learning, and the contexts in which learning occurs, all grounded in explicit 3D geometry. In doing so, we advance a geometry-grounded view of world understanding in which explicit 3D geometry shapes both learning and generalization. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35548 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35547 Effects of the Post-Heat Treatment (Furnace Cooling) on the mechanical strength and dimensional accuracy of 3D printed PEEK in FDM method Deng, Yunxiang In recent decades, the production of polymeric parts using fused deposition modelling (FDM) has gained significant attention in the field, owing to its design flexibility, low cost, and time-efficient prototyping capabilities. Nevertheless, the inherently as-built limitation constrains the performance and challenges the broader applications. To address these limitations, the post-heat treatment or annealing has long been applied as one the critical post processing techniques for enhancing the materials properties and its performance. Despite the beneficial effects of the post-heat treatment on the mechanical strength, its effect on the long-term tribological performance with the involvement of complex structures are limited. While it is often assumed that improvements in mechanical properties lead to enhanced tribological performance, tribological properties are not intrinsic material properties. Instead, they are instead dependent strongly on the specific system and operating conditions in which a material or structure has to function. Among the tribological studies, the friction-induced vibration (FIV) is a critical issue, causing unwanted noise, wear, and potential system failure. Although the proposed active or passive controls can mitigate FIV, they inevitably increase the complexity in the design and implementation of the whole system. The re-entrant auxetic structure was employed in this study as the solution, characterized by Negative Poisson’s ratio (NPR). Notably, the performance of AM-fabricated parts remains highly sensitive to the external environmental stimuli, particularly temperature. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35547 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35545 Attention Calibration for Reducing Hallucination in Large Vision-Language Models Zhu, Younan Large Vision-Language Models (LVLMs) exhibit impressive multimodal reasoning capab- ilities but remain highly susceptible to object hallucination, where models generate responses that are not factually aligned with the visual content. Recent works attribute this issue to an inherent bias of LVLMs where vision token attention map has spurious focus on certain positions, and propose to mitigate this issue by reordering visual tokens. However, we find that different LVLMs exhibit different correlations between attention and spatial position, which makes the existing static solution difficult to generalize to other LVLMs. To begin with, we investigate the attention bias introduced by image tokens through a toy experiment, in which a blank image is fed into the model to capture its position-dependent bias. We then remove this bias from the original attention map, which already leads to a substantial reduction in hallucinations. This proof of concept validates the core intuition behind attention calibration. Building on this insight, we propose Dynamic Attention Calibration (DAC), a lightweight, plug-and-play module that leverages contrastive learning to dynamically enforce positional invariance. Unlike static baselines, DAC adapts to different models and inputs in a robust and learnable manner, offering a generalizable solution to mitigate attention-related hallucinations in LVLMs. Comprehensive experiments across multiple benchmarks demonstrate that DAC significantly reduces object hallucination while improving general multimodal alignment. Our method achieves state-of-the-art performance across diverse LVLM architectures on various metrics. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35545 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35544 Design and optimisation of OpenStride an open-source inexpensive force plate actometer Yang, Yang Quantitative assessment of rodent motor behaviour is central to preclinical neuroscience, and evaluation of naturalistic movement is of particular interest. In the early 2000s, force plate actometry (FPA) was developed to track rodent subject centre-of-mass (COM) with high temporal and spatial precision, enabling varied quantifications relevant to motor performance and behaviour. While FPA was commercialised, these systems cost ~$15,000 AUD and are no longer produced. Thus, despite clear utility, particularly in movement disorders and neuropsychiatric research, FPA adoption has remained limited. In this work, we sought to develop an open-source, low-cost FPA system called OpenStride, and to characterise how its performance is shaped by hardware and signal processing choices, providing users with information needed to use it effectively across varied experimental contexts. We designed OpenStride to be able to be fabricated for ~$800 AUD using standard 3D printing and laser cutting. OpenStride achieves minimal static drift during long recordings and minimal signal displacement in response to environmental perturbation; further, it reliably tracks position and distance and can separate groups of rodents with established motor phenotypes. We characterised the influence of five key variables on measurement quality: applied load, load cell excitation voltage, mechanical damping, platform mass, and signal filtering, providing suggestions on how to optimise setup in a context-dependent manner. Our findings support the interpretation that OpenStride is capable of meaningful motor and behavioural quantification across rodent species and experimental paradigms. Its performance is influenced by modifiable parameters, and understanding these relationships allows users to configure the system to suit their specific needs. All hardware and software files have been distributed freely via GitHub, with the intent of making force plate actometry accessible to the research community. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35544 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35541 Towards Efficient High Fidelity 3D Reconstruction and Novel View Synthesis Chen, Haodong This thesis studies three-dimensional (3D) reconstruction and novel view synthesis (NVS) under tight hardware and data budgets, a common constraint in VR/AR, robotics, autonomous navigation, and cultural heritage. Existing methods often rely on multi-camera rigs, dense imagery, calibrated cameras, or heavy compute, limiting embedded and dynamic deployment. The thesis addresses this through three largely orthogonal pipeline-stage contributions: E2V for event-stream sensing, 3DLS for splatting-kernel representation, and HDGS for sparse-view depth supervision. First, E2V shows that a single neuromorphic event camera can support full volumetric reconstruction. It uses an end-to-end model to map raw events to voxel occupancy in one pass, without known intrinsics/extrinsics or multi-stage processing. A synthetic dataset of 39,739 object scans generated with an event-camera simulator is also released for training and benchmarking. Second, 3D Linear Splatting (3DLS) improves point-based radiance-field rendering by replacing the Gaussian fall-off in 3D Gaussian Splatting (3DGS) with a bounded linear kernel. This reduces blur, preserves high-frequency detail, and improves rendering speed by 30% while maintaining competitive image quality across benchmarks. Third, Hierarchical Depth-Guided Splatting (HDGS) addresses geometric inconsistency in sparse-view splatting. Its Cascade Pearson Correlation Loss (CPCL) supervises depth across scales, improving geometric accuracy and consistently outperforming prior methods when only a few views are available. Together, E2V, 3DLS, and HDGS advance 3D reconstruction and NVS toward practical use under tight energy, compute, and memory budgets. Remaining challenges include the synthetic-to-real gap, broader kernel design for splat-based rendering, and geometric consistency under extreme sparsity, occlusion, and scene dynamics. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35541 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35540 The quantitative significance of membrane lipids to the phosphorus economics of Australian native plants Liang, Grace Phosphorus (P) is an essential macronutrient that commonly limits the productivity of plants in terrestrial ecosystems worldwide. The prevalence of P-impoverished soils across the Australian continent, and throughout much of Australia’s geological history, has played a significant role in the evolution of its native flora. Many native species exhibit a wide range of traits and mechanisms that allow them to persist and thrive on some of the most P-impoverished soils, including reducing their allocation of P into membrane phospholipids. Membrane phospholipids, which account for around one-third of organic P in leaves, is a unique P compound as they can be substituted by functionally similar non-P lipids, such as galactolipid and sulfolipids. This process of substitution allows for lipidbound P to be liberated and reallocated to maintain plant growth and function. Therefore, phospholipids likely play an important role in the P economics of plants. The main focus of this thesis is on foliar membrane lipids and their quantitative significance to the P economics of native Australian flora. In particular, I investigated the contribution of membrane phospholipids in the adaptation of Australian native plants to P-impoverished soils. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35540 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35537 Using Genomic Data to Elucidate the Evolution and Historical Biogeography of the Liverwort Family Lepidoziaceae (Jungermanniales; Jungermanniopsida) Rayos, Antonio Jr Luciano Lepidoziaceae, the third-largest family of liverworts, account for almost 10% of liverwort species diversity worldwide, and there are many unanswered questions about the phylogeny and biogeography of this large and diverse family. The broad aims of this thesis were to resolve the phylogenetic relationships among the taxa within Lepidoziaceae, to propose necessary taxonomic revisions based on the resulting phylogenetic analyses, and to provide insights into the spatial and temporal distribution of the family. The thesis opened with an introduction to Lepidoziaceae, the phylogenetic studies that have been done on the family so far, and the potential of genomic data to resolve the phylogeny of the family. The remaining chapters then addressed the aims using a range of methods to analyse data sets with overlapping taxon sampling. Overall, this thesis has presented several interesting findings about the evolutionary relationships among the members of Lepidoziaceae at different taxonomic levels and provided insights into the distribution of the family across space and time. These results have taxonomic implications that can be used as the basis for taxonomic revisions. The findings from this thesis still leave many unanswered questions that can be addressed in future studies. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35537 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35536 WOMEN-FRIENDLY QURʾĀN TRANSLATIONS IN COMPARATIVE PERSPECTIVE: A FEMINIST CRITICAL DISCOURSE ANALYSIS OF LALEH BAKHTIAR’S THE SUBLIME QURʾĀN (Q.2:231, Q.4:11, Q.4:34) Elmir, Mouna This thesis investigates how English translations of the Qurʾān shape gender-conscious interpretations of verses concerning women's rights and roles. It examines Laleh Bakhtiar's The Sublime Qurʾān (2007), focusing on her translation of three Qurʾānic verses (Q.2:231, Q.4:11 and Q.4:34) that have traditionally been associated with debates on marriage, divorce, inheritance and domestic authority. The study argues that Qurʾānic translation plays a significant role in shaping ethical understandings and social attitudes towards women. It explores how alternative translation choices can challenge interpretations that have been used to justify spiritual abuse, domestic violence and patriarchal authority, while remaining grounded in the Qurʾān's ethical framework. Using an interdisciplinary methodology, the study integrates Translation Studies with Critical Discourse Analysis, Feminist Critical Discourse Analysis, Feminist Poststructuralist Discourse Analysis and Contrastive Analysis. Bakhtiar's translation is analysed in comparison with selected classical and modern Qurʾānic commentaries across textual, discursive and contextual levels. The findings demonstrate that Bakhtiar's approach is best understood as an ethically grounded and linguistically informed reinterpretation rather than a feminist ideological project. The thesis makes two original contributions to Qurʾānic translation studies by introducing the analytical category of woman-friendly translators and the concept of woman-friendly commentary as a hybrid exegetical model of tafsīr. Together, these contributions expand contemporary approaches to Qurʾānic translation and interpretation by foregrounding ethical coherence, contextual sensitivity and the Qurʾān's inclusive moral vision. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35536 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35534 Leveraging gene-by-environment interactions to identify molecular drivers of diet-induced metabolic disease Cutler, Harry Benjamin Understanding why individuals differ in susceptibility to cardiometabolic disease remains a central challenge in metabolic research. Although caloric excess is a major driver of obesity and disease risk, the threshold at which dysfunction develops varies substantially between individuals. This points to intrinsic biological mechanisms that buffer against adverse dietary environments. To identify such mechanisms, I developed a systematic experimental pipeline integrating genetic diversity across inbred mouse strains with multi-omic profiling of metabolically relevant tissues following high fat high sugar (HFHS) feeding – a perturbation that induces metabolic disease only in a subset of strains. Comprehensive metabolic phenotyping revealed that tissue-specific dysfunction was strongly shaped by genetic background, in some cases fundamentally altering which tissues were affected by HFHS exposure. Deep learning models trained solely on protein abundance in chow-fed mice accurately predicted strain- and tissue-specific responses to HFHS feeding, leading to the hypothesis that dietary stress responses are governed by the pre-existing molecular architecture of each tissue. To test this, I established a bespoke tissue-specific adeno-associated virus platform to overexpress candidate regulators identified by the deep learning models, substantially accelerating functional validation compared to conventional genetic approaches. This enabled testing of seven candidate regulators of diet-induced metabolic dysfunction. Hepatic SLC22A7 emerged as a previously unrecognised mediator of insulin resistance in both liver and white adipose tissue under HFHS conditions. Collectively, this work demonstrates that integrating systems-level molecular profiling with predictive modelling and high-throughput in vivo validation is a powerful strategy to uncouple metabolic disease from caloric excess, providing a framework for precision therapeutic discovery. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35534 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35533 Genetic Architecture and Molecular Mechanisms of Coat Colour Variation in Domestic Dogs Brancalion, Lillian Valmai Coat colour in the domestic dog (Canis lupus familiaris) is a complex trait shaped by several interacting genetic variants, regulatory mechanisms and breed-specific evolutionary history. While multiple pigmentation loci have been identified, the genetic basis of many coat colour phenotypes remains unresolved. This thesis advances knowledge of canine pigmentation genetics by identifying novel genetic variants underlying coat colour variation, resolving complex inheritance patterns, and developing practical frameworks for genetic coat colour prediction. This thesis characterises two variants associated with white-spot modification near USH2A, resolving the molecular basis and inheritance of ticking and roan phenotypes through a three-haplotype allelic series and identifying USH2A as a novel candidate pigmentation gene. Further, this thesis demonstrates that pheomelanin intensity in the Golden Retriever can be accurately predicted using a simple, biologically informed predictive model. Evaluation across multiple dog breeds highlights the importance of population-aware approaches to phenotype prediction. Finally, characterisation of the MITF-A pseudogene and reference genome misassemblies at the MITF locus facilitate more accurate genetic interpretation of this important pigmentation region. Collectively, this research expands understanding of the genetic architecture of canine pigmentation by identifying novel coat colour variants, clarifying complex genomic regions, and demonstrating how biologically informed models can translate complex traits into practical predictive tools. These findings provide valuable Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35533 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35516 Complex Individuality: The Spatial Temporal and Agential Dimensions of The Problem of Biological Individuality. Mann, Rebecca This dissertation explores how we understand and demarcate the spatiotemporally unified, cohesive wholes that we call biological individuals. There to be four main points of contention underlying the so-called problem of biological individuality: (1) Whether there is a single kind of biological individual or multiple; (2) How accounts of biological individuality address how different sorts of individuals exist in space and over time; (3) Which biological phenomena or criteria underlie biological individuality; and (4) How different concepts of biological individuality relate to one another and when they come apart. I develop a three-pronged account of biological individuality that recognises three changeable kinds of biological individuals: evolutionary individuals; organisms; and agential individuals. This is based on three core processes in the biological world: evolution, energy, and action. I begin by examining the ambiguity underlying the problem of biological individuality itself, surveying the numerous solutions to the problem. After dispelling monist approaches to biological individuality, I argue that we should treat biological individuality as an umbrella concept, under which there are multiple kinds of biological individuals based on the important ways in which biological processes form unified, cohesive wholes. I then show how the problem of biological individuality has both spatial and temporal dimensions, and that any solution must address both. I then develop a novel metabolic account of the organism, defining the organism as having a centred metabolic network, before arguing for an agential account of biological individuality, which is not coextensive with organismal or evolutionary individuality concepts. This analysis not only sheds new light on many commonly discussed problem cases, such as social insects and aggregative slime moulds, it also illuminates often-overlooked complex cases of biological individuality, including mammals like ourselves. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35516 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35515 Stand to your post: The impact of incentives on retention in the Australian Defence Force Plummer, James The Australian Defence Force has publicly stated its strategic objective to expand operational capability. An integral part of that strategy is an increase in the number of service personnel, which requires higher rates of enlistment and (or) retention. This thesis considers three aspects of military service or reward that are expected to influence the retention decisions of personnel; retention bonuses, housing and relocation. It employs data rarely available to researchers, and uses empirical methods to identify how various factors influence the retention of military personnel. This thesis is intended to further academic understanding of a relatively unexplored aspect of labour economics, namely military workforces. It is also intended to inform decision-making by policymakers on a topic of national importance. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35515 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35513 Using Machine Learning to Solve Stochastic Differential Equations Miltchinov, Nikola Dimitrov Scientific machine learning is a rapidly growing field that draws from and combines many disciplines, including computational mathematics, computer science, physics and more. In recent years, great attention has been turned towards a set of methods which allow one to employ machine learning techniques to solve differential equations. These so-called `physics-informed' methods are powerful tools which do not suffer many of the usual limitations of traditional solvers, making them incredibly well-equipped to handle those complex problems which have eluded researchers for years. Up until now, these machine learning models have focused primarily on deterministic differential systems, with little attention paid to those driven by randomness or noise. However, these stochastic problems can be just as interesting, and are by no means less important than those of the deterministic variety. The content presented in this thesis centers around the application of existing machine learning solvers to problems in stochastic calculus, something which has rarely been attempted, and which has up until now excluded a vast range of meaningful problems from consideration. Of particular interest is the Landau-Lifshitz-Gilbert (LLG) equation, which presents unique challenges when attempting to solve. We find that physics-informed solvers are not particularly well suited for dealing with problems of a stochastic nature - at least in their most basic forms - though they do have potential. Whilst there are certain methods which can somewhat bridge this gap, it will ultimately take more research before the models handling stochastic differential equations are able to catch up with those of their deterministic counterparts. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35513 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35510 Engineering DNA origami nanopores: From structural dynamics to selective molecular transport Meppat, Sreelakshmi Cell specificity is essential for effective targeted drug delivery and for minimizing off-target effects. DNA origami nanopores have been developed to interact in highly specific ways with lipid membranes, facilitating controlled membrane transport. The size, shape and chemical functionality strongly influence their in vivo fate, including endosomal escape. In this thesis a Barrel-encapsulated DNA origami nanopore (BN) nanostructure design was investigated. BN consists of a DNA origami barrel linked to a DNA nanopore via single-stranded scaffold tether. The structural properties of BN variants were validated by evaluating tether variants designed using different scaffold sequences and testing the ability to encapsulate the nanopore inside the barrel. Also, BN was tested for selective interaction with lipid membranes and for de-coupling membrane docking from membrane insertion. A Giant unilamellar vesicle (GUV) based dye influx assay was introduced and optimised to validate the membrane interactions of BN. Docking and signal responsive nanopore insertion was validated by testing different configurations of BN with a 6 helix bundle (6hb) nanopore. Toe-hold mediated nanopore switching was employed to validate nanopore activation once the BN was docked onto the membrane, triggering membrane transport of dye. The barrel docking mechanism was further modified for selective docking and competitive docking to different sub-populations of GUVs. The barrel efficiently docked onto the GUV membrane when chol-DNA handles were present on GUVs. BN docking specificity provides an approach for targeting multicellular systems and can be further enhanced through appropriate functional modifications such as aptamer or protein. The 6hb nanopore is capable of separating membrane docking from insertion, which offers a versatile platform for therapeutic cargo delivery with potentially reduced off-target effects. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35510 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35471 Macroecological processes impact soil microbial diversity and functional potential Du, Mingming Soil microorganisms are crucial for nutrient cycling and ecosystem function, yet their large-scale distribution patterns and functional responses to environmental change remain insufficiently resolved. This thesis investigates how climate, soil properties, and land-use intensification shape soil microbial communities across Australian soils, with a particular focus on community composition, ecological traits, and functional potential. By collecting soil samples at local and continental scales and applying advanced bioinformatics and machine learning, this thesis provides a clearer picture of soil microbial biogeography. The results show that broad-scale climatic constraints act as dominant filters of microbial structure and function at the continental scale, whereas soil properties, particularly pH, become the principal controls of microbial variation at the regional scale. Within this abiotic template, land-use intensification further modifies these patterns by reshaping microbial diversity, composition, and community stability, while also promoting biotic homogenisation. Beyond taxonomic patterns, the thesis shows that microbial functional attributes, including antibiotic resistance potential and growthrelated ecological strategies, are non-randomly distributed across pedo-climatic gradients. These functional patterns reveal clear ARG hotspots in climate-stable regions with agricultural activities and a higher microbial potential growth rate in higher-latitude regions. Overall, this thesis advances current understanding of soil microbial ecology by linking biodiversity, functional potential, and environmental filtering within a unified continental framework. It provides new ecological knowledge into how soil microbiomes respond to natural and anthropogenic gradients and offers a foundation for predicting microbial contributions to soil health and ecosystem functioning under ongoing environmental change. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35471 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35470 Mapping porphyry copper prospectivity from Jurassic to recent times using spatio-temporal machine learning Alfonso, Christopher Peter Copper is a crucial material for the global economy, required for the production of a vast array of modern technologies. It is likely that new copper deposits will need to be discovered; however, much of the Earth’s surface has already been explored, driving the search for these deposits to greater depths and necessitating the development of new exploration and prospectivity mapping techniques. Here we present a machine learning method for estimating prospectivity for porphyry copper – the most abundant copper deposit type. Unlike the majority of prospectivity mapping studies, our approach makes use of time-dependent datasets coupled to global plate tectonic reconstructions to quantify the likelihood that a deposit formed at a given place and time in the past. By considering the many factors which may contribute to the formation of these deposits, this method also provides quantitative information reaffirming existing porphyry copper deposit models. These deposit models often emphasise the need for thick continental crust and the subduction of large quantities of volatile material such as water or carbon to stimulate deposit formation. Correspondingly, inspection of the global or regional scale prospectivity models, through techniques such as feature importance and partial dependence analysis, indicates that input variables such as overriding plate thickness and subducting carbonate and silicic sediment volumes are strongly associated with the presence of deposits, suggesting that one or more of these factors may be prerequisite for deposits to form. These models represent a powerful new tool, both for prospectivity mapping and for furthering the understanding of the mechanisms behind the formation of porphyry copper deposits. Furthermore, the methods used to produce the models are predicted to be sufficiently generalisable to allow their application to other deposit types such as porphyry molybdenum and gold, which are also formed along subduction zones. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35470 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35469 Schism and Development: Liberalism in the Contemporary Chinese Context Zhou, Yunfan As one of the most influential political ideologies in the contemporary Chinese public sphere, liberalism not only continued to develop its theoretical tradition during the Reform and Opening-up period but also constructed a unique political vision based on its core concepts, thereby forging a powerful political consensus within civil society. However, with the alteration of the Party’s basic political line following Xi Jinping’s ascent to power, coupled with the re-consolidation of the authoritarian system, the socio political environment for Chinese liberalism has undergone a drastic transformation. Faced with new empirical and ideological challenges, contemporary Chinese liberalism has experienced profound shifts in its theoretical frameworks and discursive paradigms, rendering this intellectual tradition intensely dynamic and highly fluid in the new era. Through an exploration of the controversies and introspections among contemporary Chinese liberal intellectuals regarding critical issues in the public sphere, this thesis seeks to assess the recent developmental trajectory of Chinese liberalism—a major intellectual tradition—in the Xi Jinping era. Spanning domains from the market economy and pluralism to conservatism and nationalism, these pivotal issues form the core of liberal contemplation regarding contemporary Chinese political life. Indeed, it is through relentless introspection on these very topics that contemporary Chinese liberalism has catalyzed both its theoretical evolution and its internal schisms, ultimately articulating a political vision in the new era. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35469 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35464 Wish for the perfect card: Exploring computational metagame design in competitive strategy games Elton-Pym, Alexander This thesis explores how computational metagame design, a new emerging field, can be applied to competitive strategy games. In particular, we explore computational metagame design for the collectable card game Hearthstone, which is a valuable research platform due to the complexity of designing Hearthstone metagames. The research follows a research-through-design methodology whereby we develop three prototypes of computational tools for assisting in metagame design. These prototypes are a user interface, a simulation engine, and a metagame model. We find a number of challenges associated with the design of metagames, contribute a novel method of procedurally generating cards, and a new means of understanding dynamic metagames through modelling. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35464 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35459 Effectiveness of strengthening interventions in very weak muscles of people with spinal cord injury Chen, Lydia This thesis evaluates the effectiveness of strength training interventions for improving very weak muscles (grade 1–2) in people with recent spinal cord injury (SCI). Neurological weakness is a major consequence of SCI, limiting mobility and functional activity. Strength training is therefore central to rehabilitation, aiming to maximise strength, function, and independence. Two commonly prescribed approaches are high-dose voluntary contractions and strength training combined with electrical stimulation (ES). This thesis investigates the effectiveness of these interventions. Four projects are included: two randomised controlled trials (RCTs) and two secondary analyses. The first multi-centre RCT (n=120) examined whether 10,000 voluntary contractions over eight weeks improved strength. Participants were randomised to high-dose training plus usual care or usual care alone. The mean between-group difference was 0.4/13 points (95% CI −0.5 to 1.4), indicating little to no effect. The second multi-centre RCT (n=60) studied strength training combined with ES versus usual care. The between-group difference was 0.7/13 points (95% CI −0.7 to 2.1), below the pre-specified clinically worthwhile effect (1 point), suggesting minimal effectiveness. The third project pooled data from both RCTs to assess an impression of change scale. A moderate correlation with measured strength (Spearman’s rho = 0.5) supported its validity as a simple patient-reported outcome. The fourth project examined assessor blinding across four RCTs (n=219). Assessors were more likely to correctly guess allocation in the experimental group (OR 2.55, 95% CI 1.47 to 4.43), highlighting challenges in maintaining blinding. Overall, the RCTs showed that the studied interventions provided little to no meaningful improvement in very weak muscles after SCI. Results of the secondary analyses highlight the value of patient-reported outcomes, and challenges of maintaining assessor blinding in clinical trials. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35459 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35458 The impact of CDKN2A/B deletions in IDH-mutant astrocytomas Yuile, Alexander CDKN2A/B homozygous deletion (HoD) is associated with poor prognosis in IDH-mutant astrocytomas and is now a defining criterion for WHO grade 4 classification. However, its clinical and biological significance remains incompletely understood. We hypothesised that CDKN2A/B HoD identifies a clinically and biologically distinct subtype of IDH-mutant astrocytoma. A comprehensive literature review highlighted uncertainty regarding the impact of CDKN2A/B HoD on tumour behaviour and outcomes. We then conducted a multicentre retrospective cohort study and patterns-of-care survey. CDKN2A/B HoD was confirmed as a robust adverse prognostic marker in routine clinical practice, independent of the molecular assay used to determine deletion status. Survey responses demonstrated that deletion status influences treatment decisions. As CDKN2B HoD occurred only in tumours with CDKN2A HoD, subsequent analyses focused on CDKN2A. To investigate its biological impact, we generated p16-CDKN2A knock-in cell lines from CDKN2A-null IDH-mutant glioma models. CDKN2A loss was associated with altered growth, increased production of the oncometabolite 2-hydroxyglutarate (2-HG), and differential responses to standard therapies. RNA sequencing identified dysregulation of cell-cycle and immune-related pathways associated with CDKN2A loss. These findings were validated in patient transcriptomic datasets. We further assembled a transcriptomic cohort comparing CDKN2A HoD and non-deleted IDH-mutant astrocytomas while controlling for morphological grade, confirming that the observed differences were independent of grade. Single-cell RNA sequencing provided additional insight into the effects of reduced CDKN2A expression. Collectively, these findings demonstrate that CDKN2A HoD drives biological changes in IDH-mutant astrocytomas beyond those attributable to tumour grade alone and supports the development of therapies targeting vulnerabilities associated with CDKN2A loss. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35458 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35457 Remembering and Forgetting in Sydney: Hauntological Dialogues in Time and Place Beeke, Jane This thesis is situated within the emerging field of critical hauntology. I draw on this theoretical approach to explore how experiences of hauntedness can illuminate the entanglements of place, memory and history. Drawing on ten semi-structured interviews with individuals who reported encounters with ghosts and experiences of haunting in and around Sydney, my research examines how participants constructed meaning around these experiences. I consider how these narratives provoke broader questions around the construction of official museological memory in contrast to vernacular memories and suppressed histories. The analysis is grounded in the theoretical insights of Jacques Derrida and Walter Benjamin, with a transdisciplinary approach drawing on social psychology, archaeology and history to account for the layered dynamics of these affective experiences of place-time. My findings suggest that haunting functions not merely as a private or uncanny experience, but as a kind of knowledge that unsettles reified or official mythologies of place. In the colonised context of Sydney, these spectral encounters provoke alternative forms of cultural memory that resist erasure and open space for counter-narratives to dominant historiographies. By foregrounding the spectral as a way of knowing place, I demonstrate the potential of hauntology to reconfigure understandings of place and to challenge the politics of memory in postcolonial urban landscapes. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35457 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35456 Multimodal Learning for Organ Transplant Diagnostics: From Gene Expression to Histopathology Robertson, Harry Liam Organ transplantation remains the only solution for end-stage organ failure, yet chronic immune insults and reactive monitoring limit long-term graft survival and patient quality of life. This work develops statistical and computational methodology for the multimodal assessment of allograft dysfunction, enabling tools that detect dysfunction, characterise its mechanistic basis, and support clinical decision-making before irreversible injury accrues. The Pan-organ Resource for Molecular Allograft Dysfunction (PROMAD) compiles more than 150 studies and over 12 000 samples, providing the first multi-organ transcriptomic atlas for discovery and benchmarking of transplant biomarkers. Using PROMAD, we identify 158 genes, dominated by CXCL-chemokines and granzymes, shared across kidney, heart, liver and lung rejection, with single-cell data implicating pro-inflammatory monocytes and macrophages. Transferable Omics Prediction (TOP) converts gene-expression into ratio-based features and weights them by cross-cohort transferability, yielding biomarkers that are stable across platforms, centres and populations. An organ-agnostic, blood-based biomarker derived from PROMAD and TOP predicts biopsy-proven rejection (AUC0.81) and out performs organ-specific tests and conventional laboratory metrics in a single-centre prospective validation study. Further, to address the heterogeneity in kidney transplant histopathology scoring, we develop Renal-Agent which ensembles vision-language models to segment biopsies, score Banff lesions and generate evidence-linked reports, lifting inter-observer agreement above 0.9 across validation cohorts. Together, the aims of this thesis enable a transition from siloed, batch-sensitive diagnostics with poor reproducibility to scalable, data-driven decision-support tools. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35456 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35455 Mechanisms of viral manipulation of integrins and the TGF-β signalling pathway Gracie, Nicholas Peter Viruses manipulate host signalling to optimise replication, spread and immune evasion. The TGF-β pathway is of particular interest given its roles in epithelial integrity, fibrosis, wound repair and suppression of antiviral responses, and its dysregulation is increasingly recognised as a hallmark of severe viral disease. However, how distinct viruses engage this pathway and what consequences they have remain poorly defined. This work investigates how vaccinia virus (VACV), the live smallpox vaccine, and SARS-CoV-2, two impactful pathogens from unrelated families, exploit TGF-β signalling to promote replication and persistence. Using CRISPR-Cas9 KOs, we found that VACV activates the R-SMAD transcription factors (SMAD2 and SMAD3) independently of the canonical TβR1 kinase. R-SMADs played contrasting, non-redundant roles: loss of SMAD3 impaired replication and plaque expansion, whereas loss of SMAD2 increased cell motility and cell–cell spread. A kinase screen identified CK2 as a candidate non-canonical activator, providing a mechanism for TβR1-independent signalling. For SARS-CoV-2, the Spike glycoprotein potently induced TGF-β signalling through integrin motifs. The ancestral RGD sequence within the receptor-binding domain activated SMAD3, induced SERPINE1/PAI-1 and suppressed type I interferon responses. These effects occurred with the recombinant S protein, pseudotyped lentivirus, and authentic infection, and were abolished by an RGD mutation or the antagonist ATN-161. Despite RGD loss in Omicron beyond BA.2, the S protein retained activity through an emergent LDV site, with motif mutation and α4 antibody neutralisation implicating α4 integrins. We propose that integrin-binding motifs and TGF-β signalling are intrinsically linked and essential for viral fitness. Together, these findings reveal convergent strategies by which poxviruses and coronaviruses hijack TGF-β signalling, defining two mechanistically distinct routes and highlighting new therapeutic targets. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35455 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35454 Contribution of the Lateral Collateral Ligament on Distribution of Forces in the Canine Elbow: An Ex Vivo Study Costello, Sorcha Rebecca Objectives: To determine the effects of lateral collateral ligament (LCL) deficiency on load distribution within the joint, specifically contact force, pressure, and area between the medial and lateral compartments of canine cadaveric elbows under simulated axial loading during the stance phase. Additionally, this study aimed to evaluate whether prosthetic ligament reconstruction restores normal biomechanics within the joint. Study Design: Ex-vivo cadaveric mechanical evaluation Methods: Elbow joints (N=3) were subjected to four testing conditions using a non-constrained limb press model (condition A: intact LCL, condition B: cranial crus of the LCL transected, condition C: entire LCL transected, condition D: prosthetic LCL placed), with each condition tested in triplicate. Tactile array sensors placed in the medial and lateral compartment allowed continuous direct measurement of contact area, contact pressure and contact force. Limbs were loaded on a calibrated electromechanical testing system to 200 N under each testing condition at a 135-degree standing angle. Results: Force and contact pressure within the medial compartment increased as a result of full transection of the LCL when compared as a proportion to the lateral compartment (P = 0.023 and P = 0.045 respectively). Placement of a prosthetic LCL did not restore normal intra-articular joint contact mechanics. No significant change in contact area was observed between testing conditions. Conclusion: A LCL deficient elbow results in increased forces within the medial compartment, prosthetic ligament placement does not entirely restore normal elbow biomechanics. Prosthetic ligament placement may reduce loading of the medial compartment; however, the technique requires further development to reduce variability in loading effect. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35454 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35453 Towards Large Scale Distributed Key Generation Mai, Tiancheng Distributed key generation (DKG) is a cryptographic primitive that produces one public key and shares of a corresponding secret key among a group of distributed parties. As the basis of threshold cryptography, the classical DKG protocols are resurging due to their widespread applications in blockchain, such as MEV protection, checkpointing into Bitcoin, and more. Those applications raise a challenge of deploying DKG on a very large scale. While efforts have been made to improve DKG communication, practical large-scale deployments are yet to come due to various issues. On the other hand, theoretical challenges arise when we aim to optimise the latency while achieving near-optimal communication costs. This thesis focus on solving scalable DKG on both practical and theoretical landscape. We first investigate how to build a practical, scalable and adaptively secure DKG protocol to enable all-hands checkpointing in blockchains with weighted validators. Our Any-Trust DKG achieves (quasi-)linear computation and broadcast overhead per-node cost with the help of a common coin, against weak adaptive adversaries. The key to our improvements lies in delegating the most costly operations to an Any-Trust group together with a set of techniques for adaptive security. Our Any-Trust DKG leads to a fully practical instantiation of Filecoin's checkpointing mechanism, in which all validators of a Proof-of-Stake (PoS) blockchain periodically run DKG and threshold signing to create checkpoints on Bitcoin, to enhance the security of the PoS chain. Then, on the theoretical side, we purpose Circular Dragon, the first construction of adaptively secure DKG protocol that (i) realises optimal n/2 resilience in the synchronous network, and (ii) attains near-optimal asymptotic complexities, i.e., O(n^2 log n) communication and O(k log n) rounds. In addition, the proposed DKG protocol also produces field-element secrets to support the standard discrete-logarithm based threshold cryptosystem. Wed, 01 Jan 2025 00:00:00 GMT https://hdl.handle.net/2123/35453 2025-01-01T00:00:00Z https://hdl.handle.net/2123/35452 Socio-Ecological Determinants of Young Men’s Help-Seeking: Barriers and Pathways to Engagement Palmer, Robert Guided by the socio-ecological model, this thesis examines the determinants of young men’s help-seeking across multiple levels of influence and identifies practical strategies to support more effective and timely engagement with healthcare. Using a sequential, multi-phase, mixed methods approach, four interrelated studies were conducted. These included: a systematic review synthesising the existing evidence base on young men’s barriers and facilitators to help-seeking; secondary analyses of nationally representative survey data examining age differences in help-seeking barriers and the psychosocial determinants of these barriers for young men; and a qualitative study involving semi-structured interviews with a diverse sample of 29 young men. Together, these studies enabled a comprehensive examination of help-seeking determinants across individual, interpersonal, organisational, and societal domains. The findings clarify the unique considerations young men face when seeking help and highlight several key targets for intervention. Across studies, masculine attitudes, health literacy, service accessibility, and social contexts were consistently identified as central influences on young men’s readiness and ability to seek support. A cross-cutting insight was the important role of compassion, both self-directed and from others, in shaping young men’s engagement with healthcare and reducing perceived barriers. Taken together, this thesis provides a multi-level understanding of what holds young men back from seeking help and what can support their timely engagement with healthcare. The findings point to clear, actionable strategies, including strengthening health literacy, addressing loneliness and social disconnection, improving service accessibility, and embedding compassion within service design and public messaging. These insights offer practical pathways for improving early engagement and strengthening health outcomes for young men across the lifespan. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35452 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35444 Mesoscale Modelling of Concrete Materials under Complex Loadings Liu, Qingchen Concrete is a widely used construction material. In marine environments, its durability and mechanical performance are critical, especially regarding chloride-induced corrosion and structural behaviour under complex stresses. As concrete is a complex composite exhibiting heterogeneity across length scales from the nano- to the macroscale, this work particularly focuses on the mesoscale, as this scale allows the assessment of macroscopic performance while supporting it with microscopic evidence. At the mesoscale, the concrete can be treated as a three-phase composite of cement paste, aggregates, and the interfacial transition zone (ITZ). Aggregates, as the strongest and most impermeable phase, significantly influence performance. However, many studies simplify aggregate shapes, and controlling shape irregularity experimentally is difficult, leaving the effects of realistic aggregate morphology on concrete diffusivity and fracture insufficiently understood. To bridge this gap, this research conducts mesoscale modelling of concrete with realistic aggregate shape to provide new macro- and microscopic insights into the influence of aggregate shape on various single-physical behaviours of concrete. These include ionic diffusion related to chloride-induced corrosion, as well as concrete fracture under triaxial and dynamic loading conditions that represent complex stress environments. The findings provide a new understanding of concrete transport and mechanical behaviours, highlighting the role of aggregate shape irregularity in the diffusivity and strength of concrete. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35444 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35443 Designing for Casual Human-Robot Collaboration in Urban Public Spaces Yu, Xinyan Robots are evolving beyond their traditional roles in controlled environments and are increasingly being deployed in dynamic public urban spaces (e.g., delivery robots). To ensure smooth deployment in these settings, robots must navigate around people and respond to unpredictable situations in which they may experience operational difficulties and require human assistance. Unlike robots in relatively static environments such as laboratories or domestic settings, urban robots encounter a diverse public, most of whom are bystanders without a pre-determined intention to interact. This shifts the dynamics of human–robot collaboration, giving rise to what I term casual human–robot collaboration: forms of collaboration that emerge spontaneously during encounters. These contextual and relational shifts necessitate tailored strategies to facilitate spontaneous interactions between urban robots and bystanders. This thesis aims to develop interaction strategies that facilitate casual collaboration between urban robots and bystanders. First, an online ethnography study was conducted to identify opportunities and gain preliminary insights into how casual collaboration may emerge. The thesis then adopts a research-through-design approach, beginning with bodystorming to generate design considerations. These considerations inform design concepts evaluated in three empirical case studies, including VR lab studies and an in-the-wild field study. The studies examine three scenarios: human–robot spatial conflicts, robot–environment misalignments, and robot technological limitations. This thesis contributes to HCI and HRI in four ways: it identifies design opportunities for casual human–robot collaboration; presents design artefacts and reflections through the RtD process; provides empirical insights into how people engage with these concepts; and offers a conceptual framing for understanding casual human–robot collaboration in public spaces. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35443 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35442 THESE WALLS HAVE EYES AND EARS On architectural memory and conservation Zammit, Sarah-Jane The desire to keep alive the memory of the past is deeply human and arguably universal. However, the set of claims the conservation movement has made about buildings as witnesses to time and repositories of memory are historically specific. In the nineteenth century, romantic writers used their artistic skills to anthropomorphise buildings as key to humanity’s memory, witnesses to the passage of time and human achievement, instigating what would become a long-standing cultural tradition. This idea persists today and characterises many contemporary claims for the protection of historic places. The question remains however, why do we expect buildings to remember and do memory work? This doctoral project investigates this long-standing cultural tradition. This thesis argues that the concept of architectural witnessing relies on human perception and the human mind’s complex abilities to remember, imagine and position oneself in different spatiotemporal contexts. The research seeks to understand the cultural mechanisms at work in the nineteenth and twentieth centuries which posited buildings as the memory and history of a city, region or nation. It identifies two key historical moments of crisis around memory which has shaped claims about the role of conservation in preserving and prolonging memory - the nineteenth century c.1810-1850 and the post-WWII period in the twentieth century. Using psychology and memory science as the conceptual apparatus the thesis interrogates these two key periods. The thesis establishes that building conservation, restoration and reconstruction is analogous for the psychological apparatus of memory. Ultimately arguing that this is why we expect buildings to remember, exploring the impact this expectation has on collective memory; how society deals with material evidence and authenticity in these periods; and the creative and recreative tendencies of post-war reconstructions as re-imagined through visual media like photography and painting. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35442 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35441 Determining changes in Eucalyptus litter during decomposition Hung, Tsz Ching Christy This study demonstrated that both natural and thermal decomposition alter litter carbon and nitrogen content, affecting decomposition rates. Climate was the dominant driver, with rapid initial mass loss under warm and wet conditions. At a regional scale, decomposition was governed by initial litter quality, particularly C/N ratio; for example, Eucalyptus litter with high C/N decomposed more slowly than N-rich understorey litter. Under natural environment, N-rich litter facilitated Eucalyptus decomposition through nutrient transfer. Prescribed burning, while effective for fuel reduction, also regulates litter decomposition and nutrient dynamics. Litter collected one year after fire decomposed faster than pre-fire litter, largely due to fire-induced changes in C/N ratios. In dry sclerophyll forests, post-fire understorey regeneration further enhanced decomposition, potentially supporting forest recovery through increased nutrient availability. However, carbon and nitrogen stored in surface litter are redistributed via volatilisation and mineralisation, with implications for ecosystem nutrient dynamics. These findings highlight the need to consider litter quality, understorey vegetation, and fire regimes in fire prone dry sclerophyll forests. Traditional methods for measuring litter quality are costly and labour-intensive. Spectroscopic techniques such as visible-near-infrared (vis-NIR) and attenuated total reflectance Fourier transform (ATR-FTIR) offer efficient alternatives. Across this thesis, performance ranked ATR-FTIR > NIR > vis-NIR. While vis-NIR and NIR were effective for predicting the quality of fresh litter, ATR-FTIR provided consistently accurate predictions of carbon and nitrogen across litterfall, decomposition stages, and combustion residues. Overall, spectroscopy is a powerful complement to laboratory analysis for understanding chemical transformations in forest litter. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35441 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35436 Enhancing Graph Representation Learning: Data-Aware Advances in Graph Neural Networks Xiao, Ye Graph-structured data is a fundamental data modality across diverse domains, in which the relations among entities can be as crucial as the entities themselves. However, the irregular nature of it renders traditional Euclidean-based deep learning methods inadequate for modeling its intricate interdependent structure. Graph Neural Networks (GNNs) have emerged as the de facto paradigm for learning effective representations that capture relational information, offering favorable performance through their message-passing mechanism. Despite their development, many recent advances address different tasks primarily in a model-centric view, revolving around more sophisticated designs to enhance representation learning. Considering that the evolving demand for classification further drives models toward more complex architectures, this thesis aims to contribute to the advancement of effective representation learning by placing data at the center of methodological design. To this end, the thesis delves into various data-aware schemes to efficiently and delicately exploit the potential of the graph, thereby enhancing the graph representation learning capabilities of GNNs. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35436 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35431 Translation of Digitally Enabled Health Initiatives for Human Development Pujitha Gunawardena, Dinushika Sathsarani Access to digital technologies alone is insufficient to achieve desired improvements in low- and middle-income countries (LMICs). Historically, efforts have focused on exporting solutions from Western settings to LMICs, with limited attention to local social, economic, political, and infrastructural contexts. Such approaches result in failure to achieve expected development outcomes, frustrating local users and posing a high risk for short-lived initiatives. Moreover, they remain inadequate for realising global aspirations for human development as envisioned by the United Nations. This thesis reimagines current approaches to digital development by proposing a translational approach that offers a promising path for the design and local embedding of digitally enabled health initiatives. Chapter 2 provides a conceptual review of human development, leading to a conceptual leap that emphasises translating Information System (IS) artifacts and associated knowledge to better support aspirations for human development. In Chapter 3, a four-phase research design is developed to guide the translation-focused design of digitally enabled health initiatives. Locally anticipated outcomes in Ecuador and Papua New Guinea inform the future design of a virtual care system for their contexts through translation. Chapter 4 investigates how translation facilitates the local embedding of a digitally enabled health initiative in Sub-Saharan Africa. How actors translate IS artifacts and associated knowledge to enhance their relevance within local practices, norms, and cultures is examined. This thesis makes five key contributions: it offers a reimagined approach for digital development research; broadens the understanding of what translation entails in digital development; contextualises and extends the lens of the IS artifact; makes a methodological contribution through a translation-focused research design; and identifies interrelated outcomes of virtual care in remote settings. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35431 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35429 Capturing Students' Conceptual Change When Exploring Decimals through Dynamic Digital Representations Gorman, Amelia Kate Decimals are of great significance in the primary mathematics curriculum due to their application and use in everyday life. The purpose of this study was to investigate the effectiveness of certain dynamic digital representations in developing students’ knowledge of decimal fractions. Task-based interviews were used with six Year 4 (9-10 years old) students, that incorporated four different dynamic digital representations of decimals. Data collected via video-audio recordings were used to detect shifts in students’ attention while using the digital representations. Attention shifts were analysed using microgenetic methods to determine conceptual changes over time. Findings uncovered specific features of the digital representations that generated productive cognitive con-fusion which prompted changes in students’ understanding of decimal fractions. The unique affordances of each digital tool offered students opportunities to dynamically explore decimal concepts and could be used to enrich the teaching of decimals within the primary mathematics classroom. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35429 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35426 Confronting Plate Models with the Deep Mantle New, Thomas Christopher Mantle circulation models, numerical models of whole-mantle convection driven by tectonic reconstructions, are widely used to predict present-day mantle structure and, in turn, to evaluate the reconstructions themselves. Yet many studies compare modelled temperature fields directly with seismic tomography, overlooking nonlinearity in the temperature-velocity relationship, the strongly heterogeneous resolution of seismic imaging, and misfit metrics that saturate when anomalies do not overlap. This thesis develops a robust, physically consistent approach to evaluating competing tectonic reconstructions against tomography. I first develop a transferable comparison methodology, demonstrated with the G-ADOPT finite-element modelling library, comprising: (i) physically consistent conversion of predicted temperatures into seismic velocities; (ii) tomographic resolution operators to filter converted structures prior to comparison; and (iii) the Wasserstein metric to quantify misfit informatively even without anomaly overlap. After benchmarking G-ADOPT models across well-studied subduction regimes, I apply these tools to the collision of the Ontong-Java plateau with the Melanesian arc, where collision timing remains uncertain (~25 vs ~12 Ma). I show that physically consistent post-processing can flip qualitative inferences of slab morphology, shift inferred slab depths by tens of kilometres, change inferred sinking rates by up to 1.2 mm/yr, and laterally displace slab anomalies by up to 639 km. Benchmarking shows G-ADOPT reproduces mantle structure well where slab rollback dominates, but predicts incorrect dip polarity where trenches advance. The case study strongly supports the older collision scenario: only the ~25 Ma model reproduces slab material beneath the Melanesian arc, with 10-28% lower misfit, supporting a role for the collision in regional plate reorganisation. These results highlight the value of publishing tomography models alongside resolution operators. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35426 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35425 AI-driven Defect Detection and Multi-Agent Disaster Response for Civil Infrastructure Chen, Zhaohui Rapid and reliable assessment of civil infrastructure is essential for effective disaster response. Traditional inspection relies on manual, expert-driven surveys, which are time-consuming, resource-intensive, and difficult to scale under limited accessibility and evolving hazards. Although computer vision enables automated damage detection, most approaches focus on isolated perception tasks and lack support for system-level reasoning. This thesis develops an AI-driven framework that integrates robust local-scale perception with structured global-scale reasoning. At the local scale, efficient vision models are designed for defect detection under realistic conditions. A Transformer-based crack segmentation model with average pooling improves robustness in noisy environments, while a Robust Feature Knowledge Distillation (RFKD) framework transfers noise-resilient representations from teacher to lightweight student models for deployment. Vision Mamba architectures are further explored to improve scalability for high-resolution inspection. At the global scale, a multi-agent framework transforms perceptual outputs into actionable insights through role-based task decomposition, information integration, and collaborative reasoning, enabling coherent situational awareness with human-in-the-loop support. Experiments demonstrate improved robustness and scalability over conventional methods. Challenges remain in handling uncertainty and ensuring reliable long-horizon reasoning. Future work will focus on uncertainty-aware models and tighter integration with sensing and decision-support systems. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35425 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35423 Optimizing Large Language Models: Algorithmic Advancements and Model Design Strategies Bie, Fengxiang Large Language models have achieved remarkable performance across diverse tasks, but face two critical deployment challenges: (1) the key-value (KV) cache memory bottleneck that limits model deployment in resource-constrained environments, and (2) the sequential autoregressive generation latency that reduces inference throughput and user experience. This thesis presents two complementary contributions addressing these distinct challenges. First, CARE (Covariance-Aware and Rank-Enhanced) tackles the KV-cache memory bottleneck by converting pretrained Grouped Query Attention (GQA) models into memory-efficient Multi-Head Latent Attention (MLA) architectures. Unlike naive SVD approaches that ignore activation patterns, CARE introduces activation-preserving factorization using covariance-weighted SVD and adaptive rank allocation via water-filling algorithms. Second, Infinigram-based speculative decoding addresses inference latency by leveraging large-scale n-gram statistics to predict multiple tokens in parallel, achieving significant speedup through CPU-optimized data structures and confidence-based acceptance strategies. Experimental results on Llama-3.1-8B demonstrate that CARE achieves up to 331% relative improvement in zero-shot accuracy over baseline conversion methods while maintaining identical KV-cache footprint. Post-conversion healing fully recovers original model performance with minimal fine-tuning. Infinigram delivers significant inference speedups across various sequence lengths and batch sizes, with acceptance rates improving for longer context matches and higher-frequency patterns. This work contributes novel methodologies combining model design strategies and algorithmic advancements for efficient large generative model deployment, providing practical solutions to key memory and computational challenges without compromising model capabilities. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35423 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35422 Fast Algorithms for Fréchet-Based Similarity Huang, Zijin Polygonal curves arise naturally when recording movement data from GPS traces, animal tracking, or eye-gaze measurements. A fundamental task is to compare such curves using a formal notion of similarity. The Fréchet distance is a widely used metric for this purpose, as it respects both the geometric proximity and the traversal order of two curves. This thesis develops faster algorithms for four Fréchet-based similarity problems, each using the freespace diagram as its central tool. In Chapter 2, we study the Fréchet distance under geometric transformations, which is essential when comparing shapes regardless of their position or orientation. We present the first improvement to the decision problem for a wide class of transformations, including translations, rotations, and scalings. In Chapter 3, we consider the Fréchet edit distance, which allows inserting or deleting vertices before measuring the continuous Fréchet distance. This variant is motivated by the need for robustness against outliers and noise in real-world trajectory data. We provide faster algorithms for several edit models, improving on prior bounds by up to a factor of k · n. In Chapter 4, we study subtrajectory clustering on c-packed trajectories, a realistic model of movement data with bounded self-overlap. We present a near-linear time approximation algorithm for the subtrajectory cluster problem, circumventing the conditional cubic lower bound for general curves. In Chapter 5, we connect the partial weak Fréchet similarity to shortest paths in weighted planar regions. We construct a near-linear size approximate spanner for the 0/1/∞ weighted region problem, and apply it to obtain an approximation algorithm for the partial weak Fréchet similarity. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35422 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35421 BayeSpace - Applying Bayesian Statistics to Environmental and Ecological Phenomena Davis, Samuel Caradog Environmental and ecological phenomena are inherently complex, dynamic, and uncertain — requiring analytical tools that can generate accurate predictions while transparently quantifying uncertainty. This thesis introduces BayeSpace, a modular Python framework for applying Bayesian inference to spatiotemporal environmental and ecological modelling. It addresses key limitations in existing approaches, including fragmented workflows, steep learning curves, and model visualisation and comparison. BayeSpace integrates two core Bayesian methodologies: Bayesian Regression (BR) for problems with known functional forms, and Gaussian Process Regression (GPR) for flexible, non-parametric modelling. Built on libraries such as NumPyro and scikit-learn, it provides core classes for model definition, prior and likelihood specification, sampling, and visualisation. Key innovations include automated experiment tracking, domain generation, data simulation, kernel and transformation flexibility, and robust visualisation tools. The framework's capabilities are demonstrated through several applications. In marine cloud brightening research, BR was used to infer atmospheric dispersion parameters from real-world plume data over the Great Barrier Reef, suggesting oceanic dispersion may be more intense than terrestrial models predict. This case study illustrates BayeSpace's ability to explore complex phenomena, identify model limitations, and understand asymptotic behaviour. In species distribution modelling, GPR effectively handled spatiotemporal occurrence data, with model performance strongly correlating with total grid coverage. BayeSpace's iteration ability and flexibility were vital in designing a one-size-fits-all framework for species distribution modelling. Validation using simulated and real-world datasets confirmed BayeSpace's accuracy in linear cases while revealing challenges in highly non-linear regimes where complex posteriors affect sampler convergence. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35421 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35416 Carbon and Nutrient Interactions in Cereal-Legume Intercropping Systems: Impacts of Phosphorus Fertilization Rahman, Md Zillur Phosphorus (P) deficiency is a major constraint to agricultural productivity in low-input systems where nutrient availability limits crop growth and resource-use efficiency. Although cereal-legume intercropping is widely promoted to improve nutrient acquisition and productivity, the mechanisms by which P availability regulates nutrient uptake, biological nitrogen fixation (BNF), nitrogen competition, and belowground carbon (C) allocation remain poorly understood. This thesis investigated the effects of P fertilization on productivity, nutrient acquisition, and carbon-nitrogen interactions in cereal-legume systems through a global meta-analysis and isotope-based experiments in a P-limited soil. A meta-analysis showed that P fertilization significantly increased crop yield and N and P uptake in both monocropping and intercropping, while also improving land-use efficiency in intercrops. Experimental studies using wheat-chickpea intercrops demonstrated that P availability strongly regulated belowground C allocation and BNF. Phosphorus fertilization increased chickpea biomass, P uptake, and BNF, while reducing root C allocation, indicating a shift towards symbiotic N acquisition. In contrast, wheat maintained root C investment regardless of cropping system. A 15N-labelling experiment revealed that wheat was more competitive for nitrate uptake than chickpea, particularly under P fertilization, whereas ammonium acquisition was similar between species. Fine-root traits were important predictors of nutrient uptake under P limitation. A 13CO2 pulse-chase study showed contrasting P-acquisition strategies, with chickpea maintaining high P uptake with minimal changes in rhizodeposition, while wheat relied on increased root C allocation and rhizodeposition. These findings provide new insights into nutrient acquisition and belowground C dynamics in cereal-legume intercropping and highlight opportunities to improve P-use efficiency in sustainable low-input cropping systems. Includes publication Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35416 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35415 Causes, Mechanisms, and Mitigation of Socially-Induced Nocebo Effects Saunders, Cosette Emilie The nocebo effect refers to adverse symptoms that arise in response to treatment but cannot be explained by its active properties. Despite their major clinical, social, and economic burden, nocebo effects remain understudied. Social learning, where people learn by observing or interacting with others, is a potent pathway for nocebo effects. In healthcare, where treatment experiences can be shared face-to-face and through mainstream and social media, understanding socially-acquired nocebo effects matters. This thesis examined the causes, mechanisms, spread, and reduction of socially-induced nocebo effects. Chapter 1 introduces nocebo effects, induction pathways, and key mechanisms. Chapter 2 presents a review and meta-analysis showing that observing adverse symptoms in another person produces medium-to-large nocebo effects compared with no treatment, with effects similar to classical conditioning and larger than explicit instruction. Prior studies, however, only tested cases where model and observer had identical experiences. Chapters 3 and 4 tested whether nocebo effects arise when the model’s experience differs from the observer’s. Using a virtual reality model of cybersickness and a simulated clinical paradigm, these studies showed that social nocebo effects generalise to similar treatments and contexts. This suggests social expectations are not limited to the modelled intervention, widening the scope of harm. Chapters 3 and 5 tested ways to reduce social nocebo effects. Choice was not found to reduce socially elicited cybersickness. In contrast, positive social modelling, where side effect warnings were paired with a peer reporting no side effects, reduced symptom severity. This gives initial evidence that social learning can also counter nocebo formation. Overall, this thesis advances understanding of socially-induced nocebo effects and shows that continued work is needed to develop and test ways to reduce their burden on individuals and the wider community. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35415 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35413 Multimodal Emotion Elicitation and Recognition in Virtual Reality Kuang, Zheyuan Virtual Reality (VR) has been effectively used for eliciting emotions, yet most research focuses on the intensity of affective responses rather than on how interaction influences those experiences. To address this gap, this thesis advances a validated VR emotion-elicitation dataset through two extensions. First, we add a new high-arousal, high-valence scene and validate its effectiveness in a within-subject study (N=24). Second, we create interactive and non-interactive versions of each scene to examine the impact of interaction on emotional responses. We evaluate interaction using subjective ratings and physiological signals. Our evaluation study (N=84) shows that interaction not only amplifies emotions but also modulates them in context, supporting coping in negative scenes and enhancing enjoyment in positive scenes. Multimodal Emotion Recognition (MER) increasingly depends on fine-grained, evidence-grounded annotations, yet inspection and label construction are hard to scale when cues are dynamic and misaligned across modalities. This thesis presents an LLM-assisted toolkit that supports multimodal emotion data annotation through an inspectable, event-centered workflow. The toolkit aligns heterogeneous recordings, visualizes modalities on a shared timeline, and packages synchronized keyframes and time windows as traceable event packets. It then uses modality-specific tools and prompt templates to draft structured annotations for analyst verification and editing. Building on the dataset extensions and annotation, this thesis further investigates MER modeling approaches in VR that integrate behavioural and physiological signals from VR headsets and wearable sensors. We introduce an LLM-based Mixture-of-Experts (MoE) framework, where experts specialize in different modalities and a router assigns weights to experts for each event. The goal is to connect predictions to traceable multimodal evidence and support interpretation of affective cues in interactive VR. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35413 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35412 Learning Theory for Transformers: An Operator-Learning Viewpoint Liu, Peilin Large language models (LLMs) have reshaped the foundations of artificial intelligence research and the modes of interaction between human cognition and machine intelligence. Their influence extends further still, transforming the scientific tools through which we interrogate and model the physical world. Underlying most of these achievements and breakthroughs is a dominant architecture: the Transformer. Although the Transformer was proposed nearly a decade ago, established mathematical frameworks remain insufficient to explain the complex phenomena observed in practice with Transformer-based networks, particularly large language models. This thesis offers a principled theoretical foundation for understanding the remarkable capabilities these models exhibit, grounded in a central argument that the Transformer performs operator learning during pretraining over vast text corpora. Our analysis reveals the nature of pretraining and in-context learning mechanisms of efficient Transformer structures in an operator learning framework. Transformers maps each context distribution to a response function for queries and with more samples from context distribution, they can recover information as much as possible to get a better response function with fixed and pretrained weights without any update. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35412 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35409 If 'I am woman', what is man? British masculinity and the Women's Liberation Movement, 1970-1990 Wallhead, Emma Jane This thesis examines the theory and analysis by the Women’s Liberation Movement (WLM) about men - including their sons, husbands, fathers, friends and others - in the context of the project of liberation. The Women’s Liberation Movement (WLM) from the late 1960s has been widely recognised as naming men as oppressors in the systemic subordination of women. Beyond this, however, there has been little attention paid to the specific critique of men and masculinity that was made by this influential movement. Filling this gap, this thesis uncovers a consistent thread of thinking in the critique of men and masculinity that was sustained across campaigns and across the duration of the WLM. That thinking understood masculinity as a mechanism to assert power over others but, more explicitly, pointed to specific traits such as selfishness, hierarchical thinking, aggression and detachment as socially endorsed norms of masculinity. Rather than creating new forms of ‘masculinity’, feminists argued that, if systems of oppression were to be dismantled, gender categories such as masculinity would need to be abandoned in favour of a more egalitarian world in which women and men might coexist with a shared commitment to collective responsibility, regard for others, nonviolence and care. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35409 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35407 ESSAYS ON Strategic Control and Consumer Response: Insights from Vertical Restraints and Promotional Tools Chen, Chia-Ying This thesis examines how location shapes market outcomes through its influence on both consumer behaviour and distribution channel design. The first essay investigates the effectiveness of electronic coupons using transaction, review, and mobile GPS data. The results show that consumers located farther from a focal restaurant are more likely to leave reviews after coupon redemption. E-coupons also generate positive spillover effects by increasing subsequent review activity at nearby restaurants. These effects vary across mobility segments, with transit-oriented consumers exhibiting stronger exploratory behaviour than consumers associated with workplaces or educational institutions. The second essay examines the interaction between exclusive dealing and exclusive territories in automobile distribution channels. Using multi-level data from the Chinese automobile industry, the study finds that the two contractual arrangements operate as complementary governance mechanisms in luxury markets. Dealer size negatively affects the likelihood of exclusive dealing, particularly for luxury brands. Overall, this thesis demonstrates how spatial considerations influence firm performance, market coordination, and competitive outcomes through both consumer mobility patterns and geographic distribution strategies. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35407 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35405 Responsible Management of AI Use in Organizations: Case Studies of Strategic AI Capabilities, Risk Controls, and Knowledge Management Wang, Yichen Organizations are increasingly investing in artificial intelligence (AI), yet many still struggle to translate its technical potential into reliable, and strategically aligned organizational outcomes. These challenges have intensified as AI evolves from static predictive analytics to dynamic and opaque systems embedded in core organizational processes. Existing responsible AI approaches mainly provide high-level ethical principles. As such, an important research question remains: How can organizations achieve responsible management of AI use? To address this question, this thesis presents three qualitative case studies across different AI technologies and organizational contexts, contributing to explain how organizations can develop capabilities, governance mechanisms, and control approaches for responsible use. In particular, Study 1 examines how a social-media platform develops strategic AI capability when implementing predictive AI systems. The findings identify a cyclical capability development process that enables organizations to cultivate informed agility, controlled efficiency, and anticipatory resilience in response to evolving predictive AI outputs. Study 2 investigates dynamic learning recommendation systems across three leading social-media platforms and develops a cybernetic control approach for governing black-box AI systems through buffering, feedforward, and feedback controls. This study demonstrates how organizations maintain model performance and reliable control under conditions of continuous model evolution and learning instability. Study 3 examines a cybersecurity organization adopting generative AI for organizational knowledge management. Drawing on the SECI model of knowledge creation, this study proposes a phase-dependent AI alignment approach and identifies alignment mechanisms that support accountable, and domain aligned AI-generated knowledge outcomes. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35405 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35404 Estimating Output Gaps in Open Economies De Gorostiza, Gilliane Angela Estimating the output gap remains a critical challenge for macroeconomic policy due to data limits, reporting lags, and global shocks. This dissertation extends the Beveridge-Nelson (BN) decomposition framework across three chapters to provide more reliable, informative, and timely indicators of economic slack for emerging Asian economies and Australia, demonstrating that multivariate and mixed-frequency BN frameworks improve real-time policy decision-making. The first chapter shows that the BN filter provides more reliable estimates for emerging Asian economies than Hodrick-Prescott, Christiano-Fitzgerald, or Hamilton filters. Cyclical consumption is more volatile than the output gap, and less than one-third of GDP growth fluctuations stem from trend shocks, countering the "cycle is the trend" view. Crucially, BN estimates suffer from smaller, less frequent revisions during major economic shifts. The second chapter reveals that while traditional domestic slack measures are uninformative for Southeast Asian economies due to structural issues and informal employment, financial and external variables are highly relevant. Financial factors dominated during the Asian and Global Financial Crises, with external variables often explaining a larger share of cyclical fluctuations than domestic output. The third chapter applies a mixed-frequency framework to Australia. It finds that the labor market's intensive margin (aggregate hours worked) offers a more significant informational contribution than headline unemployment. Furthermore, the Trade Weighted Index (TWI) provides informational value nearly equivalent to the entire financial or macroeconomic sectors combined. While domestic shocks drive most Australian fluctuations, the Global Financial Crisis was largely driven by foreign shocks. Finally, a weekly TWI indicator allows for more timely updates but does not improve nowcast accuracy relative to a monthly frequency. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35404 2026-01-01T00:00:00Z https://hdl.handle.net/2123/35403 Writing a Migrant Body: Identity Predicament and Resistance in Sinophone Fiction by Chinese Migrant Women in Australia 1996-2004 Ye, Su This thesis examines the migrant identities of mainland Chinese women in 1990s Australia through Sinophone fictional works by migrant women writers, centring the migrant body as a core site of gendered, racialized, and transnational power negotiation, exploitation, and resistance. Spanning five interwoven chapters, the study traces a thematic progression from male migrants’ socio-economic precarity and bodily marginalization to female migrants’ layered quests for belonging, sexual subjectivity, marital agency, and cross-cultural solidarity. Chapter One establishes an analytical baseline via metamorphosis and gaze subversion, highlighting the vulnerable male migrant body in women’s writing. Chapter Two explores “return” as a response to rootlessness, revealing the psychological un-recoverability of the homeland and an alternative return to Eastern Buddhist-Taoist philosophy. Chapter Three examines female sexual subjectivity as emerging not through linear Western emancipation but via imitation, sacrifice, rupture, and utilitarian coping. Chapter Four analyses migrant marriages as sites of power imbalance, where the female body becomes a locus of gendered labour displacement, exploitation, and resistance. Chapter Five explores “sisterhood” from internalized misogyny to cross-racial, cross-class solidarity. The thesis reveals that the conflicting roles of the literary “body” align with the complexity of Chinese women’s migrant identities, and that “writing a migrant body” is a unique tool to expose and resist multiple pressures. It argues that Sinophone fiction by 1990s migrant women reclaims the female body from Western Orientalist stereotypes and redefines migrant identity through nuanced portrayals of resistance and agency, revealing female solidarity as a vital form of empowerment. Migrant identity and feminist consciousness are iterative processes, forged through repeated engagement with marginalization across all spheres of migrant life. Thu, 01 Jan 2026 00:00:00 GMT https://hdl.handle.net/2123/35403 2026-01-01T00:00:00Z