17837 research outputs found
Sort by
Telementoring and homeschooling during school closures:A randomized experiment in rural Bangladesh
Get PDFUsing a randomized experiment in 200 Bangladeshi villages, we evaluate the impact of an over-the-phone learning support intervention (telementoring) among primary school children and their mothers during Covid-19 school closures. Post-intervention, treated children scored 35% higher on a standardized test, and the homeschooling involvement of treated mothers increased by 22 minutes per day (26%). We also found that the intervention forestalled treated children’s learning losses. When we returned to the participants one year later, after schools briefly reopened, we found that the treatment effects had persisted. Academically weaker children benefited the most from the intervention that only cost USD 20 per child
Feminist Corporate Social Responsibility:Reframing CSR as a Critical Force for Good
Get PDFWhile corporate social responsibility (CSR) research is now impressively broad, we identify fresh opportunities at the intersection of feminist and critical analysis to reframe this field as a force for good. We focus on the epistemological grounding of CSR in its potential to understand and change how managerial activity is interpreted and influenced for progressive ends. We approach this through a reading of the debate on CSR's limited practical use, to imagine a better methodological and purposeful future for CSR. This involves a different, feminist, political and ethical stance for researchers in relation to CSR as an object, to bring CSR theory and practice into alignment to revive its sense of purpose as a driving organizational force for good through a critical, feminist CSR. Our change‐orientated approach is based on a reading of Judith Butler's notion of critique as praxis of values; it is politically aware, reflexive, and focused on the goal of good organization to address grand, often existential, challenges. We conclude by showing how this approach to CSR brings a more transparent way of analysing practice, requiring reflexive action on the part of those working with CSR initiatives both as practitioners and as researchers to co‐produce better futures
Counter-Radicalisation in UK Higher Education:A Vernacular Analysis of ‘Vulnerability' and the Prevent Duty
Get PDFThe UK Government defines vulnerability to radicalisation as, ‘the process by which a person comes to support terrorism and extremist ideologies associated with terrorist groups’. Given this relationship between radicalisation and terrorism, in 2015 the UK Government passed legislation to enhance the national capacity to pre-emptively identify vulnerable people by co opting public sector workers. This responsibility (‘the Prevent duty’) has mandated the monitoring of citizen’s behaviours based on a relationship between vulnerability, radicalisation, and terrorism that is far from concrete. Despite this, the duty is presented as a clear and actionable framework designed to support frontline workers identify vulnerability and report cases of concern. It is within this context that our paper adopts a vernacular approach to present findings from focus groups and interviews with university students and staff about their comprehension, experiences, and evaluations of vulnerability and the duty. We approach these insights as valuable (but oft neglected) instances of ‘everyday’ security knowledge and argue that they are particularly valuable in the context of a duty that co opts those within Higher Education as counter-radicalisation practitioners and subjects. Our paper argues that conceptual, operational, and normative disconnects between Government policy and vernacular insights ‘on the ground’ mean that the duty assumes an uncertain position within UKHE to the detriment to of its stated objectives
Touch-and-feel features in “first words” picture books hinder infants’ word learning
Get PDFLittle is known about the role of book features in infant word learning from picture books. We conducted a preregistered study to assess the role of touch-and-feel features in infants’ ability to learn new words from picture books. A total of 48 infants (Mage = 16.75 months, SD = 1.85) were assigned to a touch-and-feel picture-book condition or a standard picture-book condition (no touch-and-feel features) and were taught a novel label for an unfamiliar animal by the researcher during a book-reading session. Infants were then tested on their ability to recognize the label (i.e., choose the target from a choice of two pictures on hearing it named) and to generalize this knowledge to other types of pictures and real-world objects (scale model animals and stuffed animals). Infants in the no touch-and-feel condition performed above chance when choosing the target picture, whereas infants in the touch-and-feel condition did not. Infants in both conditions failed to generalize this knowledge to other pictures and objects. This study extends our knowledge about the role of tactile features in infant word learning from picture books. Although manipulative features like touch-and-feel patches might be engaging for infants, they may detract from learning. Depending on the purpose of the activity, parents and practitioners might find it useful to consider such book features when selecting books to read with their infants
A Qualitative feasibility and acceptability study of an Acceptance and Commitment-based bibliotherapy intervention for people with cancer
Get PDFSelf-directed bibliotherapy interventions can be effective means of psychological support for individuals with cancer, yet mixed findings as to the efficacy of these interventions indicate the need for further research. We investigated the experience of individuals with cancer after using a new self-help book, based on Acceptance and Commitment Therapy (ACT). Ten participants with cancer (nine females and one male, 40-89 years old) were given access to a bibliotherapy self-help ACT-based book and participated in post-intervention semi-structured interviews. Five themes were generated from reflexive thematic analysis: (1) The value of bibliotherapy (2) Timing is important (3) Resonating with cancer experiences (4) Tools of the book (5) ACT in action. The book was found to be acceptable (self-directed, accessible, understandable content, good responsiveness to exercises) and feasible (easy to use, ACT-consistent). Although not explicitly evaluated, participants' reports indicated defusion, present moment awareness, and consideration of values, as the ACT processes that contributed to adjustment, via helping them to regain control over their lives and become more present within the moment. Findings also indicate that the intervention may be best accessed following completion of initial medical treatment
The Impact of Emotionally Evocative Information on Interpreting Accuracy in a Mock Asylum Interview
Get PDFIn asylum interviews, interpreters often relay emotionally evocative information. This study compared interpreting accuracy of emotionally evocative and neutral information. Twenty-eight Arabic-English interpreters participated in a mock asylum interview held via videoconferencing. They interpreted between an English interviewer and a Sudanese-Arabic applicant who performed a scripted interview including neutral and emotionally evocative responses. Pre-interview, interpreters completed a secondary traumatic stress measure. English interpretations of the Arabic neutral and emotionally evocative responses were recorded, transcribed, and coded for interpreting errors. Emotionally evocative responses were interpreted 4% to 8% less accurately than neutral responses, which was a significant medium to large effect. Secondary traumatic stress did not moderate differences in interpreting accuracy between conditions
Video Deepfake Classification Using Particle Swarm Optimization-based Evolving Ensemble Models
Get PDFThe recent breakthrough of deep learning based generative models has led to the escalated generation of photo-realistic synthetic videos with significant visual quality. Automated reliable detection of such forged videos requires the extraction of fine-grained discriminative spatial-temporal cues. To tackle such challenges, we propose weighted and evolving ensemble models comprising 3D Convolutional Neural Networks (CNNs) and CNN-Recurrent Neural Networks (RNNs) with Particle Swarm Optimization (PSO) based network topology and hyper-parameter optimization for video authenticity classification. A new PSO algorithm is proposed, which embeds Muller’s method and fixed-point iteration based leader enhancement, reinforcement learning-based optimal search action selection, a petal spiral simulated search mechanism, and cross-breed elite signal generation based on adaptive geometric surfaces. The PSO variant optimizes the RNN topologies in CNN-RNN, as well as key learning configurations of 3D CNNs, with the attempt to extract effective discriminative spatial-temporal cues. Both weighted and evolving ensemble strategies are used for ensemble formulation with aforementioned optimized networks as base classifiers. In particular, the proposed PSO algorithm is used to identify optimal subsets of optimized base networks for dynamic ensemble generation to balance between ensemble complexity and performance. Evaluated using several well-known synthetic video datasets, our approach outperforms existing studies and various ensemble models devised by other search methods with statistical significance for video authenticity classification. The proposed PSO model also illustrates statistical superiority over a number of search methods for solving optimization problems pertaining to a variety of artificial landscapes with diverse geometrical layouts
Merging sequential e-values via martingales
Get PDFWe study the problem of merging sequential or independent e-values into one e-value or e-process. We describe a class of e-value merging functions via martingales and show that it dominates all merging methods for sequential e-values. All admissible methods for constructing e-processes can also be obtained in this way. In the case of merging independent e-values, the situation becomes much more complicated, and we provide a general class of such merging functions based on martingales applied to reordered data.<br/
Medical Image Classification Using Transfer Learning and Network Pruning Algorithms
Get PDFDeep neural networks show great advancement in recent decades in classifying medical images (such as CTscans) with high precision to aid disease diagnosis. However, the training of deep neural networks requires significant sample sizes for learning enriched discriminative spatial features. Building a high quality dataset large enough to satisfy model training requirement is a challenging task due to limited disease sample cases, and various data privacy constraints. Therefore in this research, we perform medical image classification using transfer learning based on several well-known deep networks, i.e. GoogLeNet, Resnet and EfficientNet. To tackle data sparsity issues, a Wasserstein Generative Adversarial Network (WGAN) is used to generate new medical image samples to increase the numbers of training instances of the minority classes. The transfer learning process itself also allows the building of strong classifiers by transferring knowledge from the pre-trained image domain to a new medical domain using a small sample size. Moreover, the lottery ticket hypothesis is also used to prune each transfer learning network trained using the new target image data sets. Specifically, the L1 norm unstructured pruning technique is used for network reduction. Hyper-parameter finetuning is also performed to identify optimal settings of key network hyper-parameters such as learning rate, batch size and weight decay. A total of 20 trials are used for optimal hyper-parameter selection. Evaluated using multi-class lung X-ray images for pneumonia conditions and brain tumor CT-scans, the fine-tuned EfficientNet model obtains the best brain tumor classification accuracy rate of 96% and a fine-tuned GoogLeNet model with pruning has the highest pneumonia classification accuracy rate of 81.5%.<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!