17837 research outputs found
Sort by
Curating Queer Feminist Cinema:A Conversation Between Selina Robertson and James S. Williams -- part of Special Issue 'Queer World Cinema' guest edited by James S. Williams
Get PDF
Cluster search optimisation of deep neural networks for audio emotion classification
Get PDFAutomated patient monitoring solutions greatly benefit from audio emotion classification, although the considerable variance in individual expression and interpretation of emotions poses a challenge. Current approaches often employ standard Audio Spectrogram Transformer (AST) and deep learning models such as Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN)-based networks. However, their performance can be enhanced by integrating neural architecture search techniques using swarm optimisation algorithms. In this research, we explore AST with hyperparameter optimisation for speech emotion recognition. Three deep learning architectures with optimisable -block structures and variable filter numbers, i.e. 1DCNN, bidirectional LSTM (BiLSTM) and CNN-BiLSTM, are also proposed, enabling the optimisation of network depth and width. A novel Cluster Search Optimisation (CSO) algorithm is introduced. It incorporates Cluster Centroid Search, a Cluster Distance Improvement metric and reinforcement learning to dispatch different search actions based on clustering convergence and -learning strategies, respectively. A novel Noise Tempered K-means (NTKM) clustering model is also proposed with the integration of Gaussian-based noise insertion and cluster compactness-separation measurement, to further fine-tune the cluster centriods obtained using OPTICS clustering. CSO is used for hyperparameter and architecture search for AST and aforementioned deep networks. Attention mechanisms are also integrated with CSO-optimised networks to further enhance feature learning. We evaluate the resulting models against those devised by other optimisation algorithms across the EMO-DB, SAVEE, and TESS datasets. The empirical results demonstrate that CSO-optimised AST and CNN-BiLSTM with attention mechanisms outperform other architectures and yield favourable comparison results against those from existing state-of-the-art audio emotion classification methods
Electoral Gender Quotas and Democratic Legitimacy
Get PDFGender quotas are used to elect most of the world’s legislatures. Still, critics contend that quotas are undemocratic, eroding institutional legitimacy. We examine whether quotas diminish citizens’ faith in political decisions and decision making processes. Using survey experiments in twelve democracies with over 17,000 respondents, we compare the legitimacy-conferring effects of both quota-elected and non-quota-elected local legislative councils relative to all-male councils. Citizens strongly prefer gender balance, even when it is achieved through quotas. Though we observe a quota penalty, wherein citizens prefer gender balance attained without a quota relative to quota-elected institutions, this penalty is often small and insignificant, especially in countries with higher-threshold quotas. Quota debates are thus better framed around the most relevant counterfactual: the comparison is not between women’s descriptive representation with and without quotas, but between men’s political dominance and women’s inclusion
Finding the paper behind the data:Automatic identification of research articles related to data publications
Get PDFData papers are scholarly publications that describe datasets in detail, including their structure, collection methods, and potential for reuse, typically without presenting new analyses. As data sharing becomes increasingly central to research workflows, linking data papers to relevant research papers is essential for improving transparency, reproducibility, and scholarly credit. However, these links are rarely made explicit in metadata and are often difficult to identify manually at scale. In this study, we present a comprehensive approach to automating the linking process using natural language processing (NLP) techniques. We evaluate both set-based and vector-based methods, including Jaccard similarity, TF-IDF, SBERT, and reranking with large language models. Our experiments on a curated benchmark dataset reveal that no single method consistently outperforms others across all metrics, in line with the multifaceted nature of the task. Set-based methods using frequent words (N=50) achieve the highest top-10% accuracy, closely followed by TF-IDF, which also leads in MRR and top-1% and top-5% accuracy. SBERT-based reranking with LLMs yields the best results in top-N accuracy. This dispersion suggests that different approaches capture complementary aspects of similarity (lexical, semantic, and contextual), showing the value of hybrid strategies for robust matching between data papers and research articles. For several methods, we find no statistically significant difference between using abstracts and full texts, suggesting that abstracts may be sufficient for effective matching. Our findings demonstrate the feasibility of scalable, automated linking between data papers and research articles, enabling more accurate bibliometric analyses, improved tracking of data reuse, and fairer credit assignment for data sharing. This contributes to a more transparent, interconnected, and accessible research ecosystem
Gender Dynamics in Political Leadership: The Impact of Female Party Leaders on Ideological Perceptions
Get PDFThis thesis investigates the impact of female party leadership on the perceived ideological positioning of political parties, addressing two key research questions: first, how the gender of a party leader influences voters’ perceptions of the party’s ideological stance; and second, how voters’ perceptions and expert assessments respond to leadership changes involving a transition from male to female leaders. Using original datasets documenting party leaders’ gender identities and leadership transitions across multiple countries, the study reveals significant findings. For left-wing parties, the presence of a female leader does not significantly affect the perceived ideological stance of the party. However, for right-wing parties, a female leader is associated with a notable shift in public perception of the party’s ideology further to the right–contrary to stereotypes portraying women as more progressive. Additionally, the thesis uncovers an unexpected rightward shift in expert assessments following transitions from male to female leadership in right-wing parties. This counterintuitive finding diverges from existing literature and suggests complexities in how gendered leadership changes influence expert positions. Voter assessments, by contrast, appear largely unaffected by such transitions. These findings illuminate the multifaceted ways in which gender dynamics shape public and expert perceptions of political parties, challenging traditional assumptions and contributing to a deeper understanding of the interplay between gender, leadership, and political ideology. This research offers valuable insights for scholars in political science and gender studies alike. <br/
Measuring Software Resilience Using Socially Aware Truck Factor Estimation
Get PDFContinued timely maintenance is a key aspect of project security, but typically requires in-depth knowledge of a project's code base. Truck Factor is a metric that aims to represent how vulnerable a project is to losing this knowledge through the attrition of key contributors. However, the accuracy of existing Truck Factor estimators scales poorly with project size since they tend to ignore influential team members in managerial roles, which are more common in large projects.This work proposes SNet, a novel socially aware Truck Factor estimator based on social network analysis. SNet uses network centrality measures and social signals such as GitHub Issue interactions to estimate Truck Factor and identify Truck Factor contributors. We evaluate SNet against an existing ground truth comprised of twenty-six open source projects. Our social network analysis approach achieves superior contributor classification performance (Median F1 score = 0.8) while reducing computation time by over 2x compared to state-of-the-art estimators
Exploring attitudes towards mental health difficulties and help-seeking in older adults through a stigma lens:Does self-compassion play a role?
Get PDFAcceptability and Personalisation of a Digital Psychological Intervention for Skin Conditions: A Qualitative Study of User Perspectives
Get PDFContested Illnesses: An Exploration of Experience and Understanding
Get PDFContested Illnesses (CIs) are illnesses without determinable aetiology. Despite diverse symptomologies, CIs share limited therapeutic avenues, indeterminate diagnostic protocols, and contestation over the legitimacy of the illness. This thesis explores the lived experiences of virtual symbolic community (VSC) members with myalgic encephalomyelitis (ME), fibromyalgia (FM) and morgellons disease (MD), focusing on how VSC use impacts CI patients. This research is the first sociological investigation of MD from the patient perspective and expands existing scholarship on ME and FM. It examines the applicability of epistemic injustice – a concept that addresses the inequality among possessors and wielders of knowledge – to the CI patient experience, uncertainty, and previously undiscussed forms of knowledge production. Utilising online ethnographic and semi-structured interviews, this research provides an in-depth analysis of the lived experiences of individuals with ME, FM, and MD, the influence of VSC community cultures on CI patients, and the benefits and drawbacks of VSC participation. The analysis highlights the significance of experiences of uncertainty within medical, social, psychological, future, and moral domains. Applying epistemic injustice to the CI context reveals the impact of testimonial and hermeneutical injustices, which obscure personal understanding and limit positive health outcomes. VSCs play a crucial role in creating and exchanging knowledge and unlearning processes, providing members with valuable resources for navigating their illness journey. Finally, this research underscores the importance of VSCs in supporting CI patients, highlights the drawbacks of VSC participation, and contributes to the broader discourse on patient experiences and online health communities. The collective findings illuminate the application of a final theoretical insight, to what extent ME, FM, and MD are medicalised on the three interlinked levels, and what this insight may tell us about CIs more broadly. <br/
Data sources:
Access Repository Dashboard
Do you manage Royal Holloway - Pure? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!