42337 research outputs found
Sort by
Lives and afterlives of labels: museum exhibition labels as historical sources & the science and technology exhibition labels of National Museums Scotland, 1864-1967
No full textDespite their ubiquitousness and position as crucial components in the tradition of museum display, exhibition labels have received scant attention as primary sources within the scope of the history of museums. Traditionally relegated into the camps of the practical and the pedagogical, research on museum labels has principally centered on their production and reception rather than their potential as historical sources. In response, this thesis explores the many potentials and possibilities of historical museum exhibition labels. Combining perspectives from museology, archival science, and science and technology studies, this research reframes museum labels as dynamic material and textual objects and stakes a claim for their inclusion in the work of unpacking and understanding the histories of museums and their associated objects, exhibitions, and individuals.
The principal focus of this thesis is a previously unstudied archive of over 20,000 historical exhibition labels produced between 1864 and 1967 by what would become the Science and Technology Department of the National Museum of Scotland. The archive is believed to be one of the most complete historical label collections in existence and an assemblage of this scope and scale has never been researched in detail. Collectively, the labels provide a unique opportunity to test the applicability of historical museum labels to a myriad of questions and investigative approaches, establishing them as a new line of research within the history of museums.
The primary research question driving investigation of the label archive – How may museum exhibition labels be engaged with as historical materials and texts? – is subsequently supported by explorations into the ways labels can (re)construct histories of museum interpretive and labeling practices. To engage with those lines of inquiry, the thesis contextualizes and unpacks the lives and afterlives of the archive’s labels through three distinct perspectives. First, a historiography of the long museological tradition of writing about exhibition labels maps the archive and its labels onto the broader landscape of museum label history. Second, a novel theoretical framework positions historical museum exhibition labels as intricate and complex intellectual, cultural, social, and professional products and presents new perspectives for understanding and analyzing them. Third, a set of three detailed case studies investigates distinct assemblages of the historical label collection to consider the intricate and interwoven histories between labels, objects, and the Museum through (1) the labels associated with a Boulton and Watt beam engine, (2) the labels associated with a collection of lighthouse models, and (3) the marginalia added to labels within the archive’s record books. Collectively, this thesis explores the richness, depth, and multi-dimensionality of the archive and its labels through three lenses of work: work ascribed to labels within the history of museums, work done with labels within exhibitions, and work done to labels within the archive.
As the first piece of extensive scholarship on historical museum exhibition labels, this research privileges labels by centering them within an entanglement of institutions, collections, objects, exhibitions, and individuals. Within the necessary work of museums to engage in the critical discussions and decisions around the ideologies that govern their sourcing, collecting, storing, displaying, and interpreting practices, labels have often been obscured from consideration by the shadows of charismatic objects, individuals, and organizations. This thesis reclaims space for the label and champions its place in the ongoing work on both the history and the future of museums
Engineering synthetic metabolism for enhanced cell-free protein synthesis in the PURE system
No full textThis thesis presents advancements in the Protein synthesis Using Recombinant Elements (PURE) system, a minimal, cell-free protein synthesis platform composed entirely of purified components. PURE is widely used in synthetic biology due to its defined nature, which allows for precise control over biochemical processes. However, one of the limitations of the PURE system is its relatively low protein yield and short reaction lifetime compared to lysate-based systems, which may result from inefficient ATP recycling mechanisms or the accumulation of inhibitory by-products such as inorganic phosphate.
To address this limitation, this work integrates an ATP regeneration system into PURE, utilizing a synthetic pathway that combines the activities of pyruvate oxidase, acetate kinase, and catalase. The integrated pathway regenerates ATP by converting pyruvate, phosphate, and oxygen into acetyl phosphate, which in turn rephosphorylates ADP to ATP. Remarkably, the addition of a high initial phosphate concentration (∼10 mM), which is necessary for the reaction, does not negatively impact the protein synthesis activity of the PURE system. This finding opens new possibilities for enhancing energy efficiency in cell-free systems.
Furthermore, the ATP regeneration pathway can function both independently and in combination with the existing creatine phosphate/creatine kinase (CP/CK) system, leading to significant improvements in protein production yields. The CP/CK system utilizes creatine phosphate as a substrate to regenerate ATP through the action of the creatine kinase enzyme. In particular, the combined system is capable of synthesising up to 233 µg/ml of mCherry, representing a 78% increase in yield compared to using the creatine system alone. The robustness of this approach is demonstrated through reproducible results across multiple batches of homemade PURE, and importantly, the system’s general applicability is shown through successful implementation in the commercial PURE system, PURExpress®.
Additionally, preliminary experiments show that alternative ATP regeneration pathways, such as those based on glycolysis, could also be integrated into the PURE system to further expand its capabilities. Overall, this thesis provides a systematic approach to enhancing the efficiency of the PURE system by introducing rational, modular ATP regeneration strategies. The results lay the groundwork for broader applications of PURE in areas such as cell-free synthetic biology, metabolic engineering, and the construction of synthetic cells, where controlled and sustained energy supply is crucial for functional, long-term protein synthesis
Politics of inclusive development: the marginal lens: gendered responses to disability inclusion in Ghana
No full textThe calls for inclusive development by development organisations are rooted in hostile factors. This hostility is evidenced by the omission of disability from the Millennium Development Goals (MDGs), now foregrounded by the Sustainable Development Goals’ (SDGs) emphasis on disability data disaggregation. Disability data disaggregation drives the calls for inclusion, which have resulted in the proliferation of inclusive slogans by development organisations to showcase development efforts. Yet, development organisations face the problem of refraining from disability within this inclusive development agenda. The exclusion of disability issues has significant human rights implications and incurs an economic cost ranging from 3.0% to 7.0% of countries’ Gross Domestic Product (GDP), as reported by the International Labour Organisation (ILO); however, disability is not a ‘hot’ academic and development issue. More critically, this exclusion denotes a juncture where the fundamentals of meaning falter, resulting in a state of defiance that renders disability inaccessible and overly personal, distorting inclusion into isolated and siloed strategies.
This thesis addresses 3 key questions: 1) How is disability defined and internalised, and the cultural and institutional logics shaping these understandings? 2) How is disability operationalised through both lived experience and development practice? and 3) What alternative perspectives can challenge structural avoidance and foster relational connection as a foundation for disability-inclusive development? Drawing on a mixed methods design, the study engaged 447 research participants, 63.3% women and 36.7% men, through surveys (399), interviews (28), photovoice (11), focus group discussions (9) and document analysis across 4 development organisations working on albinism, mental health, physical impairments, and women from economically poor backgrounds in Ghana. Centring the lived experiences of 234 research participants with disabilities (52.3%), this study contributes to disability and development research through an intersectional, participatory design.
Findings reveal that singular understandings of disability—whether medical, spiritual, or social—constrain inclusive development by oversimplifying complex lived experiences; however, this thesis demonstrates that these interpretations often coexist, forming a continuum of situated truths that underscores the need for a pluralistic framework in disability-inclusive development. Furthermore, people with disabilities, especially women, offer critical insights into systemic exclusion, bringing nuance and depth to inclusion efforts through pride, autonomy and sisterhood. One participant reclaims space through fashion as self-expression; another leads radio advocacy on albinism; a third reconfigures care and identity through a ‘twinning’ relationship, resisting narratives of separateness. It centres lived narratives that disrupt and expose the limitations of dominant deficit-based models, including the social model and the International Classification of Functioning, Disability and Health (ICF). Building on this and extending Mitra and Shakespeare’s critique of the ICF, this thesis shows that agency and choice are restricted when disability is understood solely through the lens of function, activity, and participation. Instead, it foregrounds personal freedoms, such as fashion as spatial and political reclamation, radio advocacy, and ‘twinning’ as an alternative care model, offering new conceptual entry points into agency, belonging, and relational identity. These narratives deepen theoretical debates by showing how people with disabilities actively resist marginalisation and create meaning in ways that are often overlooked by dominant institutional frameworks, yet are essential to achieving inclusive development.
This thesis contributes to disability and development scholarship by demonstrating that marginalisation in Ghana stems from narrow, non-gendered conceptualisations of disability embedded in partially inclusive processes. It argues that disability inclusion is essential, not optional, for development, and positions people with disabilities as critical knowledge holders in advancing equity and inclusion. Addressing the limitations of existing literature that often sidelines disability or frames it solely through discrimination, the study introduces 2 original conceptual tools: disability pluralism, which recognises coexisting meanings and the multiplicity of disability realities and truths; and disability connectors, which reframe disability as a shared, dynamic human experience shaped by time and context to de-minoritise disability. Together, these concepts advance a more inclusive and relational framework for understanding and practising inclusive development. Furthermore, the thesis highlights a critical gap between global advances in disability data disaggregation and its practical application. It reveals that disability-focused organisations often prioritise singular disability types or assume disability status and type while overlooking intersectional dimensions such as gender and multiple disabilities. In contrast, mainstream organisations focused on women or education routinely neglect disability or limit it to visible disabilities such as physical impairments, thereby constraining inclusive operationalisation.
Lastly, this thesis advances inclusive, creative methodologies that foreground marginalised voices and expose intersectional invisibility and structural exclusion, firmly repositioning disability from the margins to the core of academic inquiry and development policy
Synthesis, characterisation and applications testing of polyester-block-polyolefins
No full textBlock copolymers (BCPs) are utilised in many industries, and there is a drive to synthesise novel copolymers with properties tailored for specific applications, and to improve the sustainability metrics of both existing and new copolymers. In particular, combining distinct polymerisation mechanisms to produce BCPs of different monomer classes, such as polyester-block-polyolefins, offers new libraries of materials. Ring-opening (co)-polymerisation (RO(CO)P) can deliver a broad variety of degradable polyesters, either through epoxide/anhydride ROCOP or the ROP of cyclic esters. Reversible addition-fragmentation chain transfer (RAFT) polymerisation offers a controlled radical polymerisation route for producing polyolefins from functionalised vinyl monomers. Designing bifunctional systems containing functionality for RO(CO)P and RAFT polymerisation is a promising route to access polyester-block-polyolefins, yet this approach remains underexplored. Overall, this thesis investigates simple and robust systems capable of delivering BCPs via RO(CO)P and RAFT, and assesses the formulations and potential applications of these copolymers.
Polyesters synthesised via epoxide/anhydride ROCOP are particularly attractive, as this route offers two tuneable monomers to tailor the desired polymer characteristics. Where the limited reports of systems combining ROCOP and RAFT polymerisation often require air-sensitive conditions for the initiator synthesis and polymerisation, Chapter 2 describes the design of air-stable alternatives. Simple bifunctional alkali metal (Li, Na, K) carboxylate salts of RAFT agents were synthesised and used to prepare polyester-block-poly(vinyl acetate) BCPs by generating a macro-RAFT agent polyester. Unusually, the salts tolerate both purified and unpurified monomer feeds for ROCOP, albeit with a decrease in polymerisation control when using unpurified monomers. Successful synthesis of the target BCPs was shown, albeit with low molecular weight polyester blocks. Further investigations showed that a commercially available xanthate salt, used as the precursor to the RAFT agents used in this chapter, could be directly used as a ROCOP initiator. This generated the macro-RAFT agent polyester in fewer steps. However, in spite of the excellent activity in ROCOP, the macro-RAFT agent generated by the xanthate salt had lower end group fidelity, resulting in decreased blocking efficiency with vinyl acetate.
To overcome the challenges with ROCOP in Chapter 2, Chapter 3 instead examined ROP of ε-caprolactone (CL) to prepare poly(ε-caprolactone)-poly(vinyl acetate) (PCL-PVAc) BCPs. This approach utilised a commercially available, bench-stable organocatalyst paired with a bifunctional hydroxy-capped RAFT agent acting as a ROP initiator. The target BCP was produced cleanly without unreacted polyester when using purified CL, and was an improvement over the ROCOP-RAFT approach in Chapter 2 in terms of both end group fidelity and molecular weight. While this system remained active for unpurified CL, unreacted PCL was consequently observed during BCP synthesis. This suggests that some PCL did not have the RAFT agent incorporated when produced under these conditions.
Additional improvements for the PCL-block-PVAc copolymers were that either polymerisation could be performed first in a stepwise approach, and that the BCPs could be prepared in a one-step, one-pot reaction. Due to these synthetic improvements, the PCL-block-PVAc copolymers were taken forwards for application testing.
In Chapter 4, the potential applications of hydrophobic-hydrophilic BCPs/polymer systems in aqueous dispersions were evaluated. The BCPs produced in Chapter 3 hydrolysed to produce PCL-block-poly(vinyl alcohol). The extent of hydrolysis and the integrity of the copolymer structures were studied by NMR spectroscopy and size exclusion chromatography. These systems were then integrated into formulations and used in detailed applications testing. The overall outlook for the thesis was summarised in Chapter 5 along with the experimental and bibliography in Chapters 6 and 7 respectively
Integrating functional genomics and semi-parametric estimation to identify binding variants likely causal for altering human traits
Get PDFUnderstanding the genetic architecture of complex human traits is a central challenge
in modern genetics with applications in drug development and precision
medicine. This thesis presents methodological advancements for the discovery
of causal variants affecting human traits. These advancements are grounded in
mathematical statistics and functional genomics and supported by extensive simulations and real-world data studies using the UK Biobank.
In the first part of this body of work we introduce a comprehensive mathematical
framework for the analysis of genetic effects on traits or disease, including single
variant effects, non-linear allelic effects, and higher-order interactions. Genetic
effects are formally defined as causal estimands, yet remain difficult to identify,
reasons for which are discussed. We then construct semi-parametric estimators
for asymptotically unbiased and efficient estimation of associated statistical estimands.
Finally, we propose a network approach, based on genetic relatedness
to account for non-independent individuals. This statistical advancement is delivered
within state-of-the-art software called TarGene. TarGene is designed to
provide performant and reproducible semi-parametric estimation routines, scaling
to biobank-scale datasets, and compatible with modern high-performance
computing platforms.
In the second part, we investigate the empirical performance of these semiparametric
estimators in the context of population genetics, using UK Biobank
data. Firstly, this is done via an extensive simulation study, leveraging flexible
generative models that can adequately represent the data generating process.
Practical violations of theoretical assumptions are illustrated as well as strategies
for their mitigation. Secondly, we contrast semi-parametric estimates to published
data produced by conventional parametric models. To this end, we perform
a phenome-wide association study (768 traits) for a well-established variant
with large effect size on the body-mass index (BMI). We observe that p-values obtained
via parametric models are substantially smaller than those originating from
semi-parametric methods. The absence of overlap between some semi-parametric
confidence intervals and those originating from parametric models highlight inflated
false discovery rates due to model misspecification. In addition, for 39 traits
our method reveals non-linear allelic effects which are commonly overlooked by
current practices in linear modelling.
Finally, we propose a paradigm based on functional genetics for the discovery
of probable causal variants and the mechanism through which they act on human
traits. These variants are likely to be causal for two main reasons: (i) they are
experimentally shown to disrupt the binding of a specific transcription factor and
are thus biologically active; and, (ii) their effect on traits is modulated via transacting
variants that were associated with the same mechanism. As a pilot study,
we use TarGene to discover putative causal variants acting through the vitamin
D receptor. For these variants, a post-analysis is performed to gain more insight
into the mechanism of action.
Overall, this thesis advances the field of population genetics in three ways.
First, it provides a robust mathematical framework within which the main challenges
in the field are formally defined. Second, it addresses the statistical estimation
challenge by removing the need for parametric assumptions and delivers
an open-source state-of-the-art software. Third, it proposes a paradigm based on
functional genomics for the discovery of putative causal variants as well as the
mechanism through which they act on human traits
Disentangled representations for improved generalisation in deep reinforcement learning
Get PDFReal-world environments are diverse and unpredictable, so Reinforcement Learning
(RL) agents need to be robust to environment changes and adapt quickly. However,
RL agents often struggle to generalise to unseen changes in the environment because
they tend to overfit to the variations seen during training. This issue is especially
problematic for image-based RL, where a change in just one variable, such as the
background colour, can change many pixels in the image. The changed pixels can lead
to drastic changes in the agent’s latent representation of the image, causing the learned
RL policy to fail. Developing robust image representations that generalise effectively
is a critical challenge in RL.
In this thesis, we investigate the potential of disentangled representations to improve
generalisation in image-based RL. Disentanglement, a prominent area of research
in unsupervised learning, aims to learn representations that separate semantically
meaningful factors of variation. Despite its potential to produce robust representations,
disentanglement is relatively unexplored in the context of RL. While the
challenges of learning disentangled representations in unsupervised learning have led
to recent approaches requiring some form of supervision, we leverage the temporal
structure inherent in RL to mitigate these challenges without the need for labelled data.
We propose methods for learning disentangled representations that not only improve
generalisation in RL but also address the limitations faced by unsupervised disentanglement
methods.
This thesis introduces three novel algorithms that integrate disentangled representation
learning into RL. The first contribution, TEmporal Disentanglement (TED), uses
the temporal structure of RL to learn disentangled representations, thereby improving
generalisation to unseen changes in the environment. The second contribution, Conditional
Mutual Information for Disentanglement (CMID), learns disentangled representations
of correlated image features, enhancing the generalisation of RL policies
under correlation shifts without requiring labelled data. Lastly, Multi-View Disentanglement
(MVD) leverages multiple camera views during training to learn disentangled
representations, improving the agent’s robustness to a reduction in available cameras
Radio frequency sensing for cognitive load and neurodegeneration monitoring with AI-driven classification
No full textCognitive load is a significant early indicator of neurodegeneration. Capturing early signs of cognitive load could delay the onset of acute dementia and other neurodegenerative conditions.
Neurodegenerative diseases, such as dementia and Alzheimer’s disease, significantly impact the healthcare sector by increasing long-term care needs, raising healthcare costs, and burdening caregivers and medical resources. With an increase in the aging population, the prevalence of these diseases is expected to rise, increasing the economic burden. Due to the progressive nature of these neurodegenerative diseases, early detection
is crucial to slow down the disease progression. While conventional medical technologies can detect cerebral blood flow variations, they are not suitable for regular monitoring due to limited accessibility, high operational costs, and the need for medical supervision. This highlights the need for portable sensors that can detect cognitive load, potentially leading to early dementia detection. Portable Radio Frequency (RF) technologies have the potential to revolutionize diagnostics by providing non-invasive, cost-effective, portable and wearable devices. These wearable sensing and imaging devices could offer timely and accurate monitoring, enabling early intervention and better disease management. This proactive approach could improve patient outcomes and reduce the overall burden on healthcare
systems. This work presents portable RF sensing for multimodal detection of cognitive load and neurodegenerative diseases. The non-invasive RF sensors are designed, developed, and evaluated on artificial brain phantoms and human subjects to validate and demonstrate their efficacy. The sensing mechanism employs AI and machine learning methods for accurate classification and real-time diagnostics. Moreover, the RF sensors are utilized for portable brain imaging to monitor stroke and brain atrophy. This research provides an innovative approach to transforming mobile healthcare by offering portable imaging analysis, diagnosis, and prognosis with minimal medical supervision
Convergence of gradient-based methods for stochastic control problems in continuous time
Get PDFStochastic control problems in continuous space and time rarely admit closed form solutions. Therefore numerical methods are required to approximate optimal controls. Inspired by their recent empirical success in reinforcement learning this thesis considers gradient-based methods for stochastic control problems in continuous time. While e.g. policy iteration algorithms (PIAs) are known to
converge exponentially in appropriate settings, each step of such an algorithm entails carrying out a full static minimization problem. Gradient-based methods replace the full minimization by taking one step in the gradient direction scaled by some step size. However, their convergence is not as well understood. Taking the step size to zero leads to a gradient flow in the space of controls. This thesis proves convergence of these gradient flows in three different settings.
The first setting is an iterative scheme known as the modified method of successive approximations (MSA), approximating open-loop controls. We show that interpolations of the iterates of the modified MSA converge to a gradient flow system. We then study the convergence of this gradient flow as time goes to infinity. In the general non-convex setting we show that the gradient term converges to zero while under convexity (strong convexity) assumptions it is shown the optimization objective converges at a linear (exponential) rate.
Next, we consider a gradient flow inspired by the continuous time formulation of mirror descent. It is shown that the dynamics can be used to approximate an optimal Markov control for stochastic control problems with a bounded convex action space. When the Hamiltonian is convex, the value functions converge at a linear rate and if the Hamiltonian is strongly convex the rate is exponential.
Finally, we introduce a mirror descent gradient flow, which approximates an optimal relaxed Markov control. The flow updates the policy based on the gradient of an entropy-regularized value function. We prove that with a fixed entropy level, the mirror descent dynamics converge exponentially to the optimal solution of the regularized problem. If the entropy level decays at suitable rate along the flow it converges to the solution of the unregularized problem at a linear rate.
Proving convergence in the second and third settings outlined above is more challenging as an optimal Markov control is approximated as opposed to the first setting where an optimal open-loop control is approximated. This is because when optimizing over Markov controls the objective function is in general non-convex while in the open-loop setting convexity of the Hamiltonian is sufficient for convexity of the objective function. To overcome this lack of convexity performance differences are proved which quantify how the convexity breaks down
Investigating potential bidirectional associations between children's age at school entry and their cognitive and socio-emotional development
Get PDFWhether there is an optimal school-entry age is widely debated, and may have important implications for societal wellbeing and equity. The topic is especially relevant within the United Kingdom (UK) given that our school-entrance ages are among the youngest in the world. Notably, the last decade has seen a steady increase in the number of UK parents electing to delay or defer their child’s school-entry, which may be suggestive of wider parental concern regarding current school entrance policies. Moreover, in 2022 the Scottish National Party passed a motion to increase the school-entry age in Scotland, meaning that research into this topic is especially timely.
This thesis sought to investigate the potential bidirectional relationship between a child’s age at school-entry and their cognitive and socio-emotional development. The first study examined whether a child’s cognitive and socio-emotional functioning, as perceived by parents, influenced the decision to delay or defer their child’s school-entry. Using a longitudinal design, the second study recruited three distinct groups, namely Deferred Preschoolers, Non-deferred School Entrants and Deferred School Entrants, to examine how age and schooling influenced children’s cognitive and socio-emotional development over the academic year. A subsample of these participants took part in study three, which used functional near-infrared spectroscopy to measure prefrontal correlates of frustration and inhibitory control, which were related to behavioural indicators acquired during the longitudinal study. Finally, the longer-term effects of age were examined in study four, to establish whether age effects were present in early adolescent cognitive and socio-emotional outcomes.
Our results highlight the centrality of children’s socio-emotional development in guiding parents’ perceptions of their school readiness and the deferral decision. Deferral appears to be especially beneficial to those children with poorer socio-emotional skills, and while significant gains were evident in socio-emotional outcomes during the deferral year, brain activation during emotional regulation was less mature relative to same-aged peers who had spent the year in primary school. However, being a year older at school entry meant that deferred entrants begin school with more mature executive functioning (EF) relative to younger entrants, which could potentially enhance school engagement and progress. On the other hand, non-deferred entrants were more likely to possess mature socio-emotional skills, assessed by parent-reported questionnaires and neural data. While they begin school with less mature EF relative to older cohort members, they benefit from the cognitive stimulation of school, showing an acceleration in verbal working memory relative to peers who remained in preschool. By the end of Primary 1, age-related differences in verbal working memory were diminishing, although an age-effect continued to be seen in behavioural and neural indices of inhibitory control. However, in the longer-term, age-related differences in many of our early adolescent outcomes had disappeared, and while older children demonstrated better processing speed and proactive-control engagement, these effects were small.
Together, our findings highlight the inter-individual differences in children’s developmental trajectories and how these influence parents’ views of their child’s readiness for school. Additionally, we demonstrate how a child’s age at school entry might influence the cognitive and socio-emotional skills with which they enter school, and discuss the role of maturation and environment in developing these skills. Finally, we discuss the implications of our findings to recent and proposed changes to Scotland school-entry policies
Access Repository Dashboard
Do you manage Edinburgh Research Archive? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!