143174 research outputs found
Sort by
WhisperD: Dementia speech recognition and filler word detection with whisper
No full textWhisper fails to correctly transcribe dementia speech because persons with dementia (PwDs) often exhibit irregular speech patterns and disfluencies such as pauses, repetitions, and fragmented sentences. It was trained on standard speech and may have had little or no exposure to dementia-affected speech. However, correct transcription is vital for dementia speech for cost-effective diagnosis and the development of assistive technology. In this work, we fine-tune Whisper with the open-source dementia speech dataset (DementiaBank) and our in-house dataset to improve its word error rate (WER). The fine-tuning also includes filler words to ascertain the filler inclusion rate (FIR) and F1 score. The fine-tuned models significantly outperformed the off-the-shelf models. The medium-sized model achieved a WER of 0.24, outperforming previous work. Similarly, there was a notable generalisability to unseen data and speech patterns
Just The Way I am Wired: CEO Generation and Corporate Carbon Emissions
No full textopen access articleIn this study, we analyse the link between CEO generational experience and corporate carbon emission pollution. When differentiating CEOs by their generational cohorts, we find that firms led by CEOs from the millennial generation and Generation X emit less carbon. Alternatively, firms led by CEOs from the Boomer generation tend to have a higher carbon footprint. One of the primary channels that explains the result is media coverage of climate concerns. With recent evidence suggesting that millennials engage more with modern media than other generational cohorts, the results of the analysis suggest that this exposure translates to reduced corporate carbon emissions when climate change concern is high. The findings are robust to alternative specifications, such as difference-in-differences regression, strict sample selection criteria and propensity score matching
Universal Lessons from the Particulars of Screenwriting
No full textChapter submitted October 2025; peer-review presently ongoing. Finalised revised draft due May 30th, 2026, with the full manuscript intended to be ready by August 2026.This chapter is concerned with what makes screenwriting a unique training ground for writers. Why does the screenwriting form render often overlooked craft techniques impossible to ignore? This chapter argues that its specific demands position screenwriting as a uniquely didactic form of writing. This chapter will interrogate screenwriting’s potential as a pedagogical tool. I teach on a UK-based, undergraduate Creative Writing programme predicated upon a holistic approach to writer development; students are encouraged to experiment in all writing forms and genres, regardless of where they feel their strengths lie. Consequently, emerging screenwriters are encouraged to write poems, emerging poets are challenged to write fiction, and emerging fiction writers to tackle screenwriting. Working with poets and novelists to write their first screenplays, I have found they are often more satisfied with their screenplays when compared to work in their so-called ‘native’ forms; in many cases, the marks and feedback reflect this. Skills acquired whilst learning to write screenplays positively impact their creative practice; this chapter seeks to explore what is happening and why that might be the case
Prediction of population aging trend and analysis of influencing factors based on grey fractional-order and grey relational models: a case study of Jiangsu Province, China
No full textopen access articleBackground With the rapid development of society, China is facing an increasingly serious problem of population aging. This trend poses new challenges to the labor force structure, public medical care construction and elderly care services, forcing the government to make a series of policy adjustments. Jiangsu Province, as a region with prominent aging problems in China, has a particularly significant aging phenomenon. Against the backdrop of the Chinese government’s active response to the challenges of aging, this study conducts an in-depth analysis of the aging trend and its influencing factors in Jiangsu Province.
Methods Based on the statistical data of the total population and the aging population in Jiangsu Province from 2011 to 2023, this study employs the grey fractional-order prediction model (FGM(1,1)) to forecast the trend of the aging population and the aging coefficient in Jiangsu Province over the next decade. Additionally, grey relational analysis (GRA) based on panel data was conducted to thoroughly examine the relevant influencing factors of population aging in Jiangsu Province. The analysis identified key factors such as general public budget expenditure, health technicians, urbanization rate, and education level as being highly correlated with population aging.
Results The results of trend prediction indicate that the elderly population in Jiangsu Province is projected to continue increasing over the next decade, with the degree of aging becoming more pronounced. Additionally, GRA based on panel data reveals that factors such as general public budget expenditures and the number of health
technicians significantly influence the aging process. This suggests that public financial investment and the quantity
and quality of health technicians play crucial roles in shaping the aging trend.
Conclusions In conjunction with the analysis results from FGM(1,1) model and GRA of panel data, this study enhances the comprehensive understanding of the aging issue in Jiangsu Province. The insights derived herein offer crucial data support and a scientific foundation for both Jiangsu Province and the Chinese government to develop policies addressing population aging. Considering the anticipated future trends in aging, it is recommended that the government revise fertility policies to optimize population structure, increase investment in public finance and medical security, and promote the development of elderly care systems. These measures aim to mitigate the challenges posed by aging and achieve sustainable economic and social development
Communication Strategies for Building and Maintaining Collaborative Relations: A Sino‐American Case Study of Rapport and ‘Politeness’
No full textopen access articleThe aim of this research was to examine the communication strategies used to build and maintain collaborative relations: to identify which strategies were perceived to be rapport-enhancing, which were perceived to be rapport-undermining, and how participants responded to the latter in order to maintain rapport. The data for the study were collected during a three-week visit to the USA by a delegation of senior Chinese government officials who wanted to enhance their links and positive relations with their counterpart organisation in the USA. Various types of data were collected, the main ones being video/audio recordings of the official meetings and metapragmatic comments made by the Chinese delegates at daily evening meetings when they discussed their daytime experiences. It was found that the delegates were extremely conscious of any progress or undermining of their key goal for the visit, regularly commenting on this. They were very appreciative of the US hosts’ relaxed interaction style and for the ways in which they built common ground. Their main complaints related to the interpreter’s behaviour. The findings are discussed in relation to pragmatic research into (im)politeness, rapport management and the impact of context. Delegates seemed unaware of possible differences in interlocutors’ normative expectations associated with communicative events, especially speaking procedures and role responsibilities. The article ends by making some professional development recommendations in relation to this
An improved reference point-based evolutionary algorithm for dynamic multi-objective optimization with preferences
No full textThe file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link.Most existing studies in dynamic multi-objective optimization (DMO) focus on tracking changing Pareto optimal sets and/or the entire Pareto-optimal fronts (PFs). However, there may be scenarios where decision-makers have specific requirements or preferences (e.g., reference points) and are only interested in certain portions of the PF, known as the region of interest (ROI). To address the challenge of preference incorporation in DMO, this paper proposes an improved reference point-based multi-objective evolutionary algorithm with a preference tracking mechanism that simultaneously tracks the changing PFs and ROIs. More specifically, the parameter controlling the size of ROIs in a reference point-based dominance relationship is designed to adaptively adjust with generations for quickly searching the dynamic ROIs. Furthermore, a preference tracking method based on least squares fitting is proposed to predict preferences from historical reference points, enabling the tracking of dynamically changing ROIs. Our proposed method is compared with eight state-of-the-art algorithms on 16 benchmark problems. The empirical results demonstrate the effectiveness of the proposed method both on most test instances and in a real-world application
Disinfection of Healthcare workers’ uniforms, is domestic laundering enough and safe?
No full textAim. In the UK, domestic laundering machines (DLMs) are commonly used to clean healthcare worker uniforms, raising concerns about their effectiveness in microbial decontamination and their role in antimicrobial resistance (AMR) development. Guidance states that nurse’s uniforms should be washed at 60°C for 10 minutes or at 40°C with detergent.
Methods and results. The performance of six DLMs was assessed using bioindicators containing Enterococcus faecium and using standard and rapid 60°C washes with standard domestic non-biological detergents. The results showed that only four of the six machines achieved sufficient decontamination at 60°C during full length cycles and rapid cycles performed inconsistently. The DLMs couldn’t reach the recommended 60°C for 10 minutes especially if the DLMs were more than 4 years old. Biofilms from DLMs were also sampled and sequenced to determine the bacteria population diversity and detect the potential presence of antibiotic resistance genes. Among the bacteria detected, potentially pathogenic bacteria genera could be identified including Acinetobacter sp., Mycobacterium sp. and Pseudomonas sp. The presence of antibiotic resistance genes was also detected including efflux pump (adeF, qacG, rsmA, abaQ and abeS) , target modification mutations (VanYB, canWI and VanG) and antibiotic inhibitors (ANT 3’’ IIC). When the Staphylococcus aureus, Pseudomonas aeruginosa and Klebsiella pneumoniae were artificially and repeatedly exposed to sublethal dose of common domestic detergent, S. aureus and K. pneumoniae developed tolerance to the detergent and also cross resistance to some antibiotics (Carbapenem and Beta-lactam groups).
Conclusion. Findings suggest that current recommendations on the use of DLM for decontamination of nurse’s uniforms is not sufficient, DLMs can also be a source of AMR transmission. It is important to review laundering guidelines to ensure effective DLM performance and detergent efficacy. In addition, alternatives like onsite/industrial laundering could be considered
Decantaciones. Política y democracia cultural: Un diálogo global.
No full textLos textos reunidos en este libro corresponden a versiones actualizadas de las ponencias realizadas por las y los expositores principales de la Conferencia Internacional, Políticas Culturales y Democracia Cultural: Un Diálogo Global realizada
en el Centro Cultural Gabriela Mistral de Santiago de Chile los días 28, 29 y 30 de noviembre de 2023
Geopolitical Risk, Market Indices, and ESG Performance during Crises
No full textThe file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link.The interconnection of stock markets has been extensively examined; however, the spillover effects between conventional markets and ESG (Environmental, Social, and Governance) markets during periods of crisis remain underexplored. This study employs the Time-Varying Parameter – VAR approach to investigate the connectedness between the conventional markets of G7 countries and their ESG counterparts, along with the Geopolitical Risk (GPR) Index and Gold Index. The results indicate that the Japanese market emerges as the largest risk recipient overall, including during times of crisis. Additionally, the study reveals that market interconnectedness intensifies significantly during the COVID-19 pandemic compared to normal periods and the Russia-Ukraine war. These insights are valuable for investors and managers seeking to diversify their portfolios in times of crisis
GANCHEST: a multi-GAN-based framework for chest CXR image generation and validation
No full textopen access articleChest diseases, including COVID-19, have caused a global pandemic, resulting in a large number of deaths. In some countries, the medical system has become overwhelmed with a shortage of doctors and medical supplies, making it difficult to accommodate all patients. Deep learning has been widely used to offer smart solutions using medical images, such as chest X-rays (CXR) to identify the disease. However, it has been noticed that the majority of current studies have been based on relatively small datasets of medical images, due to the recent emergence of chest disease and the ongoing process of gathering and publishing corresponding datasets. This limited number of COVID-19 medical images may be insufficient to build robust and accurate deep learning models. To address this problem, this study proposed GANCHEST, a framework that generates Chest CXR images based on two different generative adversarial networks (GANs): the basic GAN (GAN) and the conditional GAN (CGAN). The generated images are then validated automatically through four deep transfer learning models, namely GoogleNet, InceptionV3, SqueezNet, and VGG16. The two GAN architectures' images serve as the basis for training the models, and a test set of actual chest images serves to evaluate their performance. According to the experiments, the CGAN outperformed the GAN in creating images that were more similar to the original images. Specifically, the highest classification accuracy of the CGAN achieved 92.47\% with the VGG16 model, while the highest classification accuracy of the GAN achieved 69.27\%. The GANCHEST framework, as proposed in this study, can have a wider range of applications beyond chest CXR images. It can be applied to other domains that lack datasets, such as other medical imaging modalities, where the same problem of limited dataset availability exists. The framework can be adapted to generate synthetic images that can be used to augment existing datasets and improve the performance of deep learning models. Additionally, the proposed approach of using deep transfer learning models for validation can also be applied to other fields where the need for efficient and accurate image validation arises. The GANCHEST framework is a novel approach that can be useful for several fields where the lack of datasets is a bottleneck for the performance of machine learning models