21684 research outputs found
Sort by
Multiparty session types for distributed and failure-prone systems
Distributed computer systems inherently introduce notions of concurrency, nondeterminism and possibly failure-prone communication into programs. This makes designing, and implementing, correct distributed systems difficult, and it is widely agreed that such systems require rigorous methods of verification.
Types and type systems are lightweight techniques to formal verification whereby code implementations are annotated with meta-information about the data they use. This provides certain guarantees of correctness about the usage of data within a program, and helps developers design and implement code that is correct by construction. Specific to message-passing systems, session types not only describe information about what data is, but also about how it should be communicated. Essentially, they embed communication protocols as types within a program, ensuring that well-typed code abides by the specified protocol and benefits from any properties determined on said protocol, e.g. being deadlock-free, terminating, live, etc. Multiparty session types (MPST) are a generalisation of session type theory that allows for the description of communication protocols involving two or more named participants with possibly cyclic or interleaved communication patterns.
This thesis adapts MPST to describe protocols for distributed and failure-prone systems. The contributions presented are in the form of three novel languages: MPST!, MAGπ, and MAGπ!, each using a new MPST theory targetting client-server, failure-prone and fault-tolerant systems respectively; where the unit of failure is message loss and fault tolerance is introduced through the client-server paradigm.
This work establishes foundations for designing MPST systems targetted towards building multiparty client-server and fault-tolerant programs. For the first time, MPSTs are explored within programs designed to run over failure-prone networks, where messages are susceptible to loss, unbounded delay and total reordering. The construct of replication is shown to be expressively significant for building multiparty client-server systems. Lastly, it is demonstrated how combining replication, nondeterministic timeouts and standard MPST constructs, results in an expressive type system capable of modelling—and verifying properties on—fault-tolerant multiparty programs
The ballad of Motel Fresno: reimagining narrative poetry through music and performance
Abstract not currently available
Biphotons for nonlinear imaging
The harnessing of the nonclassical properties in photon pairs generated by the process of Spontaneous Parametric Down Conversion (SPDC) to enhance two-photon interactions has long intrigued researchers. Their use in light-matter interactions has been proven to improve efficiency either by increasing the rate of generation or by suppressing noise. Namely, the processes of Entangled Sum Frequency Generation (eSFG) and Entangled Two Photon Absorption (eTPA) have been shown to produce more photons when stimulated with entangled photons than the classical versions of the effects. This has only been effective so far at very low intensities, which, alongside the inability to reproduce previous experimental results, has caused debates about the usefulness of the effect. In the presented work, I aim to shed light on the topic by employing a carefully crafted source of entangled photon pairs to improve two-photon effects. I attempt to demonstrate, through a series of experiments, that the quantum enhancement can be observed at an order of magnitude higher than previously thought possible. To solidify the results, a direct comparison of the quantum effect was conducted by investigating the performance of the phenomenon when it is stimulated by classical radiation from a laser. Our results confirm a quantum improvement in a regime previously considered impossible, and they enabled us to examine the ramifications of the effect on imaging. By doing so, we consider real-life microscopy applications, and therefore, we can set limits to the effect
English language instructors’ perceptions of equity and its practical implementation
This study examines the perspectives of English language instructors regarding equitable instruction in relation to adult community college English language learners. Equity in education is an important goal to improve student outcomes and redress historical injustices. However, research on equity in English language learning in the United States has primarily been conducted in American kindergarten through twelfth grade (K-12) settings, and when addressed in higher education, is usually related to institutional outcomes as a whole.
This study considered equity related to the diversity of the needs and goals of adult English language learners. It sought to understand how English language teachers working in programs that serve students of various needs and goals define equity and how they implement equitable instruction. Participants were recruited from Washington State Community Colleges, specifically from programs which serve international and domestic/resident students, who may have very different needs and goals, in the same program. The study provides more insight into how instructors define equity and serve students of diverse needs and goals. The design was qualitative and interpretivist and employed semi-structured interviews and a focus group to collect data. Braun and Clarke’s (2022a) reflexive thematic analysis was used to analyse the results.
Six main themes were developed through the data analysis. The study concluded that equitable instruction is made possible through the intentional efforts of instructors to facilitate access to content, learning, and cultural capital (theme 1), knowing students through centring the individual (theme 2), and prioritizing community and belonging (theme 3). The other themes highlight the factors that influence participant understanding of equity and their pursuit of providing equitable instruction. They are the instructor’s attitude of service (theme 4), the impact of political and institutional factors (theme 5), and equity as a work in progress (theme 6). These findings can help instructors develop equity-mindedness and better serve all students to realize their needs and goals
Bioengineering 3D leukaemic niches to develop stem cell therapies
Abstract not currently available
Advancing non-destructive testing with CMUT-based ultrasound imaging system: from submerged materials to industrial applications
Capacitive Micromachined Ultrasonic Transducers (CMUTs) have emerged as a powerful alternative to conventional piezoelectric transducers for ultrasonic sensing and imaging, offering significant advantages in bandwidth, scalability, and CMOS process compatibility. However, their widespread deployment in non-destructive testing (NDT) and underwater sensing has been hindered by the reliance on bulky laboratory instrumentation, limited electronic integration, and inconsistent performance in submerged environments. This thesis presents the complete design, fabrication, and experimental validation of a miniaturised, low-power CMUT-based ultrasonic imaging and measurement system, developed to enable compact, field-deployable, and energy-efficient NDT applications.
At the core of this work is a custom-designed front-end electronic platform that integrates signal generation, amplification, acquisition, and digital processing within a single compact architecture. The printed circuit board (PCB) developed in this study provides tunable DC biasing, programmable pulse generation, and wideband amplification through dedicated analogue signal paths. The system incorporates a field-programmable gate array (FPGA) for deterministic timing control, high-speed sampling, and real-time signal processing using 14-bit, 125 MSPS analogue-to-digital converters. This integration eliminates the dependence on conventional laboratory hardware such as oscilloscopes, Bias-T, and waveform generators, achieving full system autonomy with dramatically reduced size, cost, and power consumption. Extensive experimental investigations were conducted to optimise CMUT excitation parameters, including the effects of pulse width (10–80 ns), frequency (2.5–5 MHz), polarity, and DC–AC signal interplay. The results reveal that the CMUT achieves optimal performance near 3 MHz with pulse widths above 30 ns, providing strong signal amplitude and reduced harmonic distortion. The removal of the Bias-T network further enhanced transmission efficiency and improved the signal-to-noise ratio. The system was validated through a series of underwater and NDT experiments employing both pulse–echo and pitch–catch configurations, demonstrating high measurement repeatability with standard deviations below 0.12 mm and coefficients of variation under 0.01%. The developed platform was subsequently extended into a fully functional ultrasonic imaging system, featuring FPGA-based A-, B-, and C-scan processing integrated with a motorised 3D scanning stage repurposed from an Ultimaker printer. Imaging experiments on submerged stainless-steel and aluminium specimens successfully identified geometric features, surface defects, and embedded structures with submillimeter accuracy, confirming the capability of the proposed system for underwater material inspection and defect detection.
In summary, this thesis delivers four primary contributions: the development of a compact, miniaturised PCB for CMUT biasing and signal control; an FPGA-based digital architecture for deterministic ultrasonic timing and acquisition; systematic CMUT characterisation and performance optimisation under submerged conditions; and the successful implementation of a portable CMUT-based NDT and imaging system. Collectively, these innovations advance the state of the art in ultrasonic engineering by demonstrating that CMUT-based systems can achieve high precision, reliability, and autonomy in real-world inspection and underwater applications
Decoupled electrolysis using silicotungstic acid as a redox mediator for zero-carbon hydrogen production
Water electrolysis for hydrogen production has substantial potential to address the current energy crisis while mitigating environmental pollution. However, achieving truly green hydrogen production requires using materials better suited to intermittent power generation, as existing systems pose serious concerns over gas mixing and cell component degradation. This thesis examines the concept of decoupled water electrolysis, utilising silicotungstic acid (H4SiW12O40) as a redox mediator in a flow-cell system to generate hydrogen. Decoupled electrolysis offers exceptional flexibility by enabling the separate production of hydrogen and oxygen at different times and rates, reducing gas crossover issues to a minimum.
Chapter 1 explores the green hydrogen production route in a net-zero world and the various methods employed in green hydrogen production. Furthermore, we introduce the concept of decoupled water electrolysis, an emerging approach for green hydrogen production via electrochemical processes, and discuss how it can potentially address some of the challenges associated with renewable-driven hydrogen production.
Chapter 2 provides background information on the theory underlying the experimental techniques used throughout this study. Various electrochemical and analytical methods were employed to monitor changes in potential, current, charge passed, and the electrode surface during the reduction of the redox mediator.
Chapter 3 highlights the application of decoupled water electrolysis in a flow system using silicotungstic acid as the redox mediator by assembling two electrochemical cells for hydrogen production (mediator re-oxidation) and oxygen production (mediator reduction) while applying a range of commercially relevant current densities (0.05–1.35 A/cm2 ) to monitor the liquid mediator behaviour as it was circulated continuously between the two cells as hydrogen was produced.
In Chapter 4, regenerated cellulose dialysis membranes were employed in the oxygengenerating cell, in order to compare the resulting electrochemical system with the one from Chapter 3 (which used only perfluorinated membranes). A comparative analysis of the membrane (before vs after electrolysis), current density, and the decoupling efficiency (%) obtained in this section was conducted.
Chapter 5 contains overall conclusions and discusses future work. It provides a summary of the work and suggestions for future research
Development of an interdisciplinary decision support tool to inform mangrove restoration and conservation in Vietnam’s deltas
Abstract not currently available
Novel surface termination for advanced diamond electronics
Diamond is an interesting semiconductor for high-power and high-frequency devices due to its ultra-wide bandgap, high carrier mobility, and superior thermal conductivity. However, traditional doping is limited by the deep energy levels of most impurities, which limit carrier activation at room temperature. As a consequence, despite its exceptional intrinsic properties, relatively few reproducible and high-performance diamond field-effect transistors have been reported in recent literature. This motivates the exploration of alternative doping strategies, such as hydrogen-terminated diamond with surface transfer doping.
Hydrogen-terminated diamond provides a conductive two-dimensional hole gas (2DHG) with low activation energy and relatively high carrier mobility, making it highly attractive for electronic devices. At the same time, the surface conductivity of H-diamond is highly sensitive to surface chemistry, oxide interfaces, and fabrication processes, posing significant challenges for achieving stable and controllable device operation.
In this work, different surface terminations have been explored for negative electron affinity (NEA) and positive electron affinity (PEA), to clarify their respective roles in enabling surface transfer doping or suppressing surface conductivity. Subsequently, the behaviour of different contact metals on H-diamond was examined, with emphasis on their ability to form reliable ohmic contacts. The influence of different oxide layers and deposition methods on H-diamond transfer doping was studied, revealing that thermal ALD HfO₂ can enhance the 2DHG by promoting transfer doping, whereas electron-beam–deposited Al₂O₃ with prior in-situ annealing effectively suppresses surface conductivity without degrading the hydrogen termination.
Building on these findings, accumulation-channel hydrogen-terminated diamond MOSFETs were successfully fabricated using an optimised and reproducible process flow. The devices exhibit normally-off, enhancement-mode operation with an Ion/Ioff ratio of 10⁷, achieving drain current densities exceeding 35 mA/mm at room temperature.
These results demonstrate a viable pathway towards stable and controllable diamond MOSFETs, addressing key technical barriers that have limited progress in the field. A stable and reproducible Au-based fabrication process was established for hydrogen-terminated diamond devices, providing a robust contact platform for the demonstrated enhancement-mode MOSFETs
Investigation of the possibilities for host modulation therapy for periodontal treatment
Background: Periodontitis is a globally prevalent chronic inflammatory disease characterized by immune dysregulation and alveolar bone loss. Conventional therapies, while effective in many cases, often fail to achieve complete resolution, particularly in patients with systemic comorbidities or aggressive disease phenotypes. This has led to increasing interest in Host Modulation Therapy (HMT), which targets the underlying immunopathogenesis rather than the microbial component alone.
Aim: This thesis investigates the potential of HMT in periodontal care, focusing on bonemodulating cytokines and the RANK/RANKL/OPG axis. It seeks to evaluate the feasibility of repurposing denosumab, a monoclonal antibody against RANKL used in osteoporosis, as a therapeutic adjunct in periodontitis.
Methods: Two research questions were addressed. First, salivary concentrations of OPG, RANKL, and inflammatory cytokines (IL-1β, IL-6, IL-8, TNFα) were analysed in systemically healthy individuals with varying periodontal status (health, gingivitis, periodontitis) before and after treatment. Second, an exploratory clinical study assessed periodontal changes in a patients initiating denosumab therapy. Biomarker quantification was performed using multiplex immunoassays, and clinical inflammation was assessed using the Periodontal Inflamed Surface Area (PISA) metric.
Results: OPG was variably detectable across health and periodontal disease states, while RANKL levels were consistently below detection thresholds. IL-1β showed a consistent reduction following treatment, suggesting its potential as a marker of therapeutic response. In the sample studied there were no significant correlations between cytokine levels and PISA. The denosumab study was limited by recruitment challenges and evolving prescribing guidelines, but provided valuable insights into the complexities of studying biologic agents in dental populations.
Conclusion: HMT represents a promising adjunctive strategy in periodontal therapy, particularly for high-risk individuals. While current evidence supports the biological plausibility of targeting host pathways, further research is needed to validate targets, optimise delivery systems, and establish long-term clinical efficacy. The integration of personalised medicine and biomarker-guided care may enhance the future role of HMT in periodontal practice