21684 research outputs found
Sort by
Making the invisible visible: an intersectional examination of disability on the early modern stage
No full textAbstract not currently available
Machine learning in asset pricing and portfolio optimization
Get PDFIn the rapidly evolving field of finance, asset pricing and portfolio optimization are facing challenges due to technological advancements, shifting economic landscapes, regulatory changes, and increased complexity in financial markets. This thesis contains three essays that explore the use of advanced machine learning approaches in financial advising and asset pricing. The first essay improves robo-advisors’ performance by combining reinforcement learning (RL) with importance sampling that focuses on rare events, leading to better investment outcomes. The second essay employs inverse optimization to estimate investors’ risk aversion under normal and disaster conditions, and then optimizes portfolios based on the learnt risk aversion by deep RL. The third essay proposes a framework for asset pricing that uses neural networks to model nonlinear pricing kernels and includes considerations of environmental, social, and governance (ESG) factors in explaining cross-sectional asset prices. Details of the three essays are summarized below:
Robo-advising under rare disasters Robo-advisors provide automated portfolio management services to investors, and their growth has been unprecedented in the past few years. However, empirical evidence shows that roboadvisors underperformed during the recent COVID-19 pandemic. This may be because rare disasters are highly unlikely to occur and yet have a huge impact on financial markets. Our study develops a novel computational framework to improve the performance and robustness of robo-advising in the presence of rare disasters. It integrates RL with importance sampling. Instead of sampling the transition probability from a ground-truth probability distribution, we sample it from a proposal distribution, where the event of interest occurs more frequently. The proposed algorithm is validated by data covering the 2008 financial crisis and the COVID-19 pandemic, showing superior performance over benchmarked methods. The estimated quarterly return of the robo-advising portfolio using the optimal policy of the proposed algorithm is 0.512%, significantly higher than both the benchmarked policy and the average quarterly return, which are-0.639% and-14.55%, respectively. This improvement is attributed to targeted learning about rare disasters, enabling robo-advisors to reduce exposure to risky assets. The proposed algorithm is model-free and reduces the variance of value estimates through importance sampling.
Risk aversion and portfolio optimization for robo-advising We develop a novel framework for learning investors’ risk aversion using low-resolution data, a common issue arising from short trajectories recording investors’ portfolio choices, particularly during disaster events. Furthermore, the observed portfolio choice is often affected by behavioural biases. Our approach combines online inverse optimization with deep RLto simultaneously estimate risk aversion and determine optimal investment strategies under both normal and disaster states. Utilizing real mutual fund data, we demonstrate that our algorithm’s risk aversion estimation converges asymptotically to the optimal risk aversion during the learning process. Critically, based on the learned risk aversion and trained deep RL model, we show that robo-advisors adopting our approach can effectively tailor investment strategies to suit investor risk aversion under varying market conditions, outperforming traditional funds. This highlights the potential for our framework to enhance investment decision-making and better represent investor interests in both stable and volatile market environments.
Nonlinear pricing kernels via neural networks This study proposes a nonlinear pricing kernel approximated through neural networks, addressing limitations of traditional linear models, which capture linear relationships and are prone to overfitting when applied to the factor zoo. The proposed model specification test examines the validity of the nonlinearity assumption of the pricing kernel. Through optimal neural network selection, our findings reveal that a one-layer neural network significantly reduces quadratic pricing errors, indicating its superior pricing performance compared to deep neural networks. Moreover, the role of ESG variables in asset pricing, particularly within the extensive range of factors, remains underexplored. The significance test designed for neural networks shows that ESG variables are significant in asset pricing
Circulating microRNAs as diagnostic biomarkers for endocrine hypertension
No full textAbstract not currently available
Electroencephalograph predictors of central neuropathic pain in subacute spinal cord injury
No full textAbstract not currently available
The humanitarian Babel. A comparative study of Catholic relief organisations’ assistance to refugees
No full textAbstract not currently available
Education of a socially just citizenry: towards a philosophy of anti-capitalist education
Get PDFThis thesis examines theories and philosophies of social justice, linking them to a conceptual approach to citizenship education and the development of what I term a socially just citizenry. I analyse the dialectic of justice-oriented dispositions and the social structures that enable or thwart the development of such dispositions at the level of the citizenry. The aim is to provide conceptual clarification of how education can strengthen the knowledge base and social arrangements required for the practice of justice-oriented citizenship.
It is commonly asserted that education is fundamental to the realisation of social justice, yet the role of education in this regard can lack conceptual clarity in theory, policy and practice. Social justice is a contested term with a variety of meanings and implications that change over time. Similarly contested is the role of schools, educators, pedagogies and curricula in relation to social justice. Additionally, it is recognised that the education system itself can perpetuate social inequalities. Particularly relevant to these concerns is the role of citizenship education and yet there is an absence of literature that seeks to conceptualise citizenship education in direct relation to social justice and the reproduction of a socially just citizenry.
This thesis represents a philosophical exploration of the role of education in the reproduction of such a citizenry. I problematise an individualistic approach to citizenship education that seeks to produce responsible citizens while failing to adequately address issues of social justice. I propose and discuss the philosophical conjunctures of social justice and social power, drawing on radical philosophies of resistance to injustice and abuses of power through the lens of racial capitalism. The relationship of these conjunctures to the practices of citizenship is evaluated and presented as the theoretical basis for a socially just citizenry.
I begin by constructing a philosophical approach to the understanding of social justice, addressing difficulties of its definition and validity as an educational aim. I show that social justice is not a straight-forward goal but a continuous and problematic process that is best characterised, I argue, as resistance against injustice. Within this struggle I add a multidimensional theorisation of social power. I discuss theories of citizenship, differentiating citizenship as status (for example, legal status or nationality) and citizenship as practice (for example, engagement in democracy). Following this, I demonstrate the conceptual potential of the socially just citizenry to contextualise and intervene in contemporary educational debates.
While a citizenry is commonly understood as a group of citizens, I build on the perspective that understands citizenry as a wider body politic. I propose that a socially just citizenry is a dispositional and systemic orientation towards social justice that transcends national boundaries. This requires the education of citizens who are critically informed and emotionally invested in social justice, with the capacity for resisting injustice. I argue that a socially just citizenry embodies the economic, cultural, political and affective levers of democratic power that enable social transformation. Education of a socially just citizenry, as understood here, is an intersectional, relational, decolonial and anti-capitalist endeavour.
This analysis informs my argument of how justice-oriented citizenship can be developed through education. I identify critical activism as an approach to justice-oriented citizenship, proposing it as an educational imperative. I conclude that an education of a socially just citizenry is best understood within the broader project of anti-capitalist transformation. Finally, I identify routes forward for future investigations and applications in educational practice
Pursuing clarity of purpose and generalizable research practices for mental health apps and recommendations
Get PDFMental health issues are a diverse and widespread problem. Whilst effective treatments and solutions exist, numerous obstacles make it difficult to deliver them to the people who need them. One promising way to lower access barriers is through mental health apps, a ubiquitous, often inexpensive solution. Furthermore, through the existence of recommender systems, delivery of content within the app can be personalized, allowing for the personal tailoring of a widely available resource. Yet whilst there have been some promising results in the field regarding their efficacy, commercial deployment is outpacing the supporting science. Many apps are untested, there are very few unifying methodological frameworks in place and a shallow understanding of mechanisms of change. The research that does exist is comprised mostly of exploration and testing of novel algorithms or solutions leading to little coherence between studies and generalizability of findings is largely unknown. Whilst, to some extent, this is to be expected in early-stage research, it would be far more beneficial in the long run to pursue foundational improvements now rather than later.
This thesis aims to address these issues, to shore up the foundations of research into mental health apps and recommender systems, to identify generalizable practices and pursue a deeper understanding of how we can enable positive mental health change. The current work focuses on establishing the link between engagement and mental health outcomes, as research in the field has a tendency to make assumptions regarding the beneficial effects of increased engagement. Through a systematic review of recommender systems in the mental health context and three studies, we evaluate how recommendations can be tailored towards both outcomes, how we can increase congruence in research by clear, goal-oriented definition of variables, and whether academic research translates to real world effects. The current work investigates the influence of financial incentivization on engagement and mental health outcomes, analyzes a commercial dataset gathered over several years and explores the relationship between different character traits, behaviors, facets of engagement and short-term and long-term mental health outcomes across a number of domains. In the final chapter the disparate threads will be brought together, presenting broadly applicable recommendations for how researchers can structure individual research to generate more cohesive value in the field as a whole, and suggest how future research may continue to pursue a stronger foundational understanding of mental health change
Power and resistance: exploring the conceptions and experiences of gender and career aspirations of young people in Bangladesh
Get PDFThe thesis investigates the conceptions of gender and career aspirations among young people in Bangladesh who participated in a gender-sensitive programme provided by an NGO against the backdrop of Sustainable Development Goals (SDGs) 4 and 5. The thesis adopts a qualitative methodology collecting data from the Southeast part of Bangladesh. This study aims to understand how the discourses of gender and career aspirations are shaped by other socio-patriarchal discourses in Bangladesh and how these young people negotiate their positions.
In 2015, the UN General Assembly proposed 17 goals to achieve global sustainable development as part of the 'Agenda 2030' (UN, 2015). SDG 4 is to "ensure inclusive and equitable quality education and promote lifelong learning opportunities for all", and SDG 5 is to "achieve gender equality and empower all women and girls" (ibid). As a UN member state, commitment to achieving the SDGs significantly shifted Bangladesh's education focus from quantity to quality (BANBEIS, 2022). However, quality inclusive education is a significant challenge for Bangladesh (Rahman, 2021; GED, 2022), and the primary evidence in Bangladesh around inclusive quality education is based on quantitative studies with scattered investigations (Asadullah, 2016; Golam and Kusakabe, 2018; Biwas and Biswas, 2020). In this study, I argue that it is necessary to understand the conceptions and experiences of young people in relation to gender and educational aspirations to achieve inclusive and equitable quality education for all (Davies, 2003; Francis, 2006).
The research design of this study draws on qualitative feminist research utilising photo-elicitation and semi-structured interviews as data-collection methods that create a scope for the participants to share their experiences in their own words (Denzin and Guba, 2011; Coleman, 2016). The study consists of twenty female and male participants based on their self-identification. However, this study primarily focuses on the conceptions and experiences of the female young people and includes male perspectives to enrich the findings. Theoretically, the study adopts a feminist poststructuralist stance with Foucauldian theories and feminist ideas around power, discipline, and resistance to analyse how female young people are coerced, disciplined, and even punished to conform to dominant discourses of gender and career aspirations. The study also brings the spotlight on how these female young people negotiate and exercise agency to make alternative discourses possible.
The study contributes to knowledge creation by employing a feminist poststructuralist analysis of power/knowledge and discourses in educational research with a focus on gender in a South Asian context. It is a valuable example of a Foucauldian study with a feminist stance exploring patriarchy, violence, and inequality that is considered marginal in Foucault's work (Ramazanouglu, 1993). Thus, the study furthers the Foucauldian debate. The research also creates knowledge by uncovering ways discourses influence female and male young people differently. It also discusses how power/knowledge is continuously constructed and re-constructed making space for alternative discourses. The findings can support teachers to understand young people's conceptions of gender and career aspirations to make schools, classrooms, and playgrounds more gender inclusive
Sparse intrinsic Gaussian processes in complex constrained domains with application of Bayesian optimisation
Get PDFAs scientific research advances, more and more data are no longer limited to traditional Euclidean space, but extend to spaces with more complex geometric structures, such as complex constrained domains and Riemannian manifolds. Riemannian manifolds are increasingly being recognized as an important tool in data analysis and machine learning due to their widespread use in multiple scientific fields and in real-word contexts. For example, lakes can be modeled as manifolds to better understand their geographic structure and dynamics in environmental studies. In order to model such manifolds in real world situations, an increasing number of statistical tools are developed for estimation over a manifold. When considering regression on manifolds, inspired by the success of Gaussian Processes (GPs) in Euclidean spaces, this thesis aims to provide novel tools in order to efficiently and accurately estimate surfaces using GPs tailored for manifolds.
Traditional GPs typically use kernels that rely on Euclidean distance to define the covariance between data points on the target surface, such as the radial basis function (RBF) kernel. Traditional GPs cannot be directly applied to manifolds due to their failure to accurately capture the underlying structure, especially in the presence of gaps and complex boundaries. The heat kernel describes the heat diffusion on the manifold, which reflects the manifold’s geometric properties, but only specific manifolds have closed-form expressions. Intrinsic GPs proposed in [114] use the transition density of Brownian motion (BM) on the manifold to approximate the heat kernel, thereby capturing the manifold’s intrinsic geometric characteristics and enabling more accurate regression on manifolds. However, Intrinsic GPs face issues near boundaries due to resampling BM paths when crossing the boundary, causing inaccurate predictions near the boundary. According to the definition of the Neumann boundary condition, the BM path should be reflected when it crosses the boundary. This thesis proposes a "reflection" method to address this issue, leading to more accurate predictions at the boundary.
Additionally, Intrinsic GPs are constrained by the computational complexity of simulating BM paths, especially on large-scale or highly complex manifolds, which make them highly computationally intensive. This thesis investigates the feasibility of sparse methods in Intrinsic GPs, which use inducing points as intermediaries to facilitate information transmission from training points to test points, aiming to simplify the computational complexity without sacrificing inference accuracy. This thesis first proposes Sparse Intrinsic GPs using a Deterministic Inducing Conditional approach (SI-GPDIC), which is straightforward to implement and computationally efficient; however, it is sensitive to the location of a small number of inducing points. The Sparse Intrinsic Gaussian Process using a Deterministic Training Conditional approach (SI-GPDTC) is then proposed, which is less sensitive to the location of inducing points, achieving a balance between computational efficiency and inference precision. Considering approximating the true posterior distribution with a simpler, more tractable distribution by minimizing the divergence metric between them, this thesis develops the Sparse Intrinsic Gaussian Process with Variational Inference (SI-GPVI), a powerful tool for regression on complex manifolds. Graph GPs, which utilize the graph Matérn kernel on the undirected graph constructed from the manifold, and Traditional GPs, which directly use the Euclidean distance-based RBF kernel, are employed for comparison with the three Sparse Intrinsic GPs developed in this thesis. The performance of the proposed methods is demonstrated using three examples: the 2D U-shape, the 3D Bitten-torus, and the real-world dataset of the Aral Sea, with SI-GPVI performing particularly well.
Finally, motivated by the success of Bayesian optimisation (BO) in Euclidean space, this thesis proposes novel approaches to construct Intrinsic BO on manifolds, building upon previous research. The proposed GPs (introduced earlier in the thesis), serve as surrogate models in the BO approach, providing the acquisition function the probability of improvement (PI), with accurate information about the underlying manifold structure. Benefiting from the surrogate models’ ability to capture the structure of manifolds, the proposed BO algorithms—Intrinsic BO with DTC and Intrinsic BO with VI—achieve better results compared to Graph BO, based on Graph GPs, and Traditional BO, based on Traditional GPs. Among them, Intrinsic BO with DIC shows unstable performance due to its predictive variance providing inaccurate uncertainty when estimating points that are far from inducing points, whereas Intrinsic BO with VI demonstrates particularly strong performance, excelling in both accuracy and efficiency
Techniques for subtle mid-air gestural interaction using mmWave radar
Get PDFUsers need to be able to interact with mid-air gesture systems in ways that are efficient, precise, and socially acceptable. Subtle mid-air micro gestures can provide low-effort and discreet ways of interaction. This thesis contributes techniques for recognizing and utilizing subtle mid-air gestures with millimeter wave radars, a rapidly emerging sensing technology in human-computer interaction.
The first contribution focused on the problem of addressing a system. By analyzing the frequency components of various hand motions, subtle activation gestures were identified which produced high-frequency signals through deliberate, rhythmic movements. A novel activation gesture recognition pipeline was then developed using frequency analysis to recognize these gestures and ignore incidental hand motions. Tested across three types of sensors, the pipeline demonstrated robust performance in recognizing subtle high-frequency activation gestures and producing zero false activations for broad hand motions. Further improvements were also explored to enhance robustness to reduce false activations during activities like typing, writing, and phone usage.
The second contribution focused on recognition of subtle gestures from mmWave radar data using deep learning. A new dataset was developed, capturing the temporal dynamics and motion patterns of 10 different subtle gestures from 8 users with a mmWave radar. Multiple neural network architectures were trained and evaluated using the dataset, achieving a high recognition accuracy of 90%. The results demonstrated that hybrid neural networks combining convolutional and recurrent layers can effectively recognize subtle gestures from mmWave radar signals and generalize across different users.
The final contribution progressed from offline evaluations to practical, real-time assessments. The neural network models were integrated into prototype applications that enabled real-time subtle gesture interactions for tasks such as selecting photos and adjusting media playback. A user study demonstrated significant improvements in task completion, accuracy, and user experience compared to traditional macro gestures. The findings suggest that subtle gestural interaction, enabled by mmWave radar sensors, signal processing, and deep learning, can significantly enhance usability of virtual interfaces
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇