159370 research outputs found
Sort by
Evaluating explanation performance for clinical decision support systems for non-imaging data: a systematic literature review
No full textPurpose
This review investigates the effectiveness of Explainable AI (XAI) in machine learning (ML)-based clinical decision support systems (CDSS) using non-imaging data, focusing on explanation quality, clinical decision-making, user trust, and usability. It highlights clinician and patient perspectives to assess XAI's role in enhancing transparency and real-world adoption.
Method
A methodological and usability-focused systematic review using Web of Science, Scopus, IEEE Xplore, PubMed, Cochrane Library, and ACM Digital Library was conducted using keyword combinations, such as “XAI” AND “CDSS” AND “Evaluation metrics” AND “User study OR Evaluation study".
Results
The review identified an increase in multidisciplinary XAI healthcare research since 2023, with applications spanning intensive care, oncology, neurology, and clinical decision support systems. Studies commonly employed mixed-method evaluations, combining technical metrics (e.g., accuracy, fidelity) with human-centred assessments (e.g., trust, usability). Trustworthiness, interpretability, and transparency emerged as key XAI properties; however, aspects such as patient involvement, explanation usability, and clinical integration remain underexplored. Findings highlight ongoing challenges in balancing explanation faithfulness with user plausibility, and in aligning explanations with clinical reasoning and workflows.
Conclusions
The study highlights the importance of striking a balance between technical fidelity and human interpretability, achieved through human-centred evaluation frameworks that incorporate both objective and subjective metrics to enhance the real-world applicability of XAI tools. Future research should focus more on human-centred AI/XAI frameworks and real-world evaluations that prioritise multi-stakeholder collaboration to enhance clinical decision support, improve diagnostic accuracy, and enable personalised care without compromising clinician expertise or patient safety
CO2 activation on metal- and non-metal-doped goldene sheets: A DFT study
No full textGoldene is a novel two-dimensional monolayer (2D) of gold atoms arranged in a hexagonal close-packed crystal lattice. In this study, we have employed density functional theory (DFT) calculations to investigate CO2 activation on doped goldene sheets (G_X, where X represents the dopant). We targeted three distinct categories for the doping process: Alkali metals (AMs = Li, Na, K, and Rb), alkaline earth metals (AEMs = Be, Mg, Ca, and Sr), and non-metals (NMs = C, O, and Se). Our study reveals that Be doping leads to the most robust structure among all considered G_X sheets, with a binding energy (Eb) of −4.73 eV, whereas AEM dopants are found to bind more strongly with goldene compared to AM dopants. No dopant was able to induce a gap between the energy bands, and the goldene preserves its metallic character and flat geometry even after doping. The CO2 molecule adsorbs on AM-doped goldene sheets with a high adsorption energy (Eads) but remains linear after adsorption. However, in the case of AEMs, CO2 chemically adsorbs on the G_Be sheet and becomes activated by bending (136.28°). The presence of NMs (C and O) enhances the chemical activity of goldene towards CO2 conversion
Charged up with peace in mind: Unraveling the factors of range anxiety among norwegian electric vehicle drivers
Get PDFRange anxiety and its associated factors are crucial for designing effective solutions to enhance the Electric Vehicle (EV) user experience. The present study is based on a survey of 1,005 EV drivers in Norway. It aims to provide insights into the perception of range anxiety in a country with well-developed electric mobility. We contribute to the state-of-the-art by exploring the concept of range anxiety based on eleven potential explanatory variables describing EV adopters’ driving experience, charging-related user experience, psychological characteristics, and socio-demographics. Our analysis showed that the longer the EV driving experience, the greater the perceived ease of charging and charging flexibility, the lower the range anxiety will be. Further, stronger control beliefs over EV's charging state and willingness to take risks relate to lower range anxiety. Additionally, being younger and being female are associated with lower range anxiety among EV drivers. However, perceived helpfulness of charging guidance, technology openness, and settlement size of EV drivers are not significantly associated with range anxiety. Our study highlights the importance of providing situation- and user-oriented charging information, a smooth and easy-to-understand charging process, and the need to support EV drivers who are less experienced, older, and routine-based
Learning Fairer Representations with FairVIC
No full textMitigating bias in automated decision-making systems, particularly in deep learning models, is a critical challenge due to nuanced definitions of fairness, dataset-specific biases, and the inherent trade-off between fairness and accuracy. To address these issues, we introduce FairVIC, an innovative approach that enhances fairness in neural networks by integrating variance, invariance, and covariance terms into the loss function during training. Unlike methods that rely on predefined fairness criteria, FairVIC abstracts fairness concepts to minimise dependency on protected characteristics. We evaluate FairVIC against comparable bias mitigation techniques on benchmark datasets, considering both group and individual fairness, and conduct an ablation study on the accuracy-fairness trade-off. FairVIC demonstrates significant improvements (≈70%) in fairness across all tested metrics without compromising accuracy, thus offering a robust, generalisable solution for fair deep learning across diverse tasks and datasets
Beyond Economic Growth: Rural–Urban Integration and Citizens' Perceptions of Social Justice in China
Get PDFCan integrating rural and urban systems foster a stronger sense of justice in society? Since 2002, China has advanced rural–urban integration to promote balanced development and reduce inequality. However, its impact on how citizens perceive social justice remains underexplored. Using survey data from 28 provinces, this study examines how three dimensions of rural and urban integration—economic, quality of life, and social development—relate to perceived social justice. The findings reveal a striking contrast: while integration in social development is positively associated with perceptions of justice, integration in the economic and quality of life domains is negatively associated. Further analysis suggests that economic integration has failed to deliver equitable income distribution (result justice), which may contribute to public dissatisfaction. In contrast, integration in social development has enhanced both welfare provision (result justice) and social mobility (opportunity justice), improving individuals' perceptions of fairness. These findings suggest that although rural and urban integration has the potential to enhance perceptions of justice, its impact differs across reform domains. Promoting inclusive social development, particularly through welfare provision and reforms that improve social mobility, may be essential to fostering a fairer society during China's ongoing transition
Back-contact perovskite solar cells: progress, challenges and the future
No full textBack-contact solar cells are a novel type of device architecture that have been used in combination with both silicon and perovskite semiconductors. Since 2016, the efficiencies of perovskite back-contact solar cells have risen from 6.5% to over 11%, with a wide variety of fabrication methods and device structures employed. This Spotlight on Applications article aims to highlight the benefits of back-contact solar cells over more routinely used planar counterparts and reviews recent progress made using perovskite materials. In particular, we show how the use of substrates patterned with embossed microgrooves presents a facile route to back-contact solar-cell scale-up, with this approach recently used to create perovskite solar-cell modules consisting of hundreds of serially connected grooves having a stabilized efficiency of up to 12.8%. We conclude by discussing the challenges required to drive this technology toward high-volume manufacture and commercialization
The children may not be the problem: evidence of acceptance and enjoyment of higher fibre breads from choice architecture studies in school breakfast clubs
Get PDFLow fibre consumption is endemic in the United Kingdom (UK). Replacing refined staples with wholegrain starchy staples could increase fibre consumption. School food contributes to children’s nutrition and establishes eating norms, presenting both a public health opportunity and challenge. It could be a policy lever to effect generational dietary change and influence health outcomes, and goal-strategic public sector procurement. However, this policy lever is under-exploited. Despite efforts by non-governmental organizations and individual schools, school food has not been conferred high value or status in national food policy and often does not provide children with a diet rich in healthy, high-fibre foods. In the H3 project, we trialled interventions to increase the fibre content of school breakfast, with particular focus on replacing white bread—a UK breakfast staple—with higher fibre breads. Here we review this work, providing insights from the food industry and children’s perspectives. A key outcome was that child preferences were not a major barrier. When provided with higher fibre foods, children ate and enjoyed them. This suggests that simple food policy levers could significantly reduce the approximately 6 g average ‘fibre gap’ in UK school children’s intake, for example by mandating that all bread served in schools be at least a ‘source of fibre’. Larger scale trials should be conducted, ideally as part of universal school breakfast pilots
Multiwavelength Characterization of Optical Wireless Communication in Complex Water-Filled Pipe Environment
No full textThis paper presents an in-depth investigation of optical wireless communication through water-filled PVC pipelines using high-brightness light-emitting diodes (HB-LEDs) operating at visible wavelengths: 475 nm (blue), 528 nm (green), 583 nm (yellow), and 625 nm (red). Simulations were conducted in Ansys Zemax OpticStudio using ray-tracing techniques and Bidirectional Scattering Distribution Function (BSDF) models to evaluate the effects of surface roughness, interface reflection, and wavelength-dependent absorption. A custom experimental setup was developed using a 375 mm long, 50 mm diameter PVC pipe and a Thorlabs S121C photodiode sensor to validate the simulation. Optical power was measured under five water fill conditions (0%, 25%, 50%, 75%, and 100%). Results show that the greatest transmission loss occurs at the 50% water level, where multiphase scattering dominates, with experimental power decreasing to −11.82 dBm at 583 nm (yellow). Full immersion improves transmission, with recovered power levels up to −2.3 dBm at 475 nm (blue). Absorption coefficients were calculated using the Beer–Lambert Law, with peak values exceeding 0.09 cm⁻¹ at 50% fill. Simulation results aligned with experimental measurements within 1–2 dB, validating the model’s reliability. These findings support the development of adaptive gain control strategies and wavelength-optimized optical links for autonomous robotic inspection in submerged or semi-submerged pipeline environments
Investigating the DNA binding properties of two new RuII(dppz) complexes incorporating the facially coordinated ligands tris(pyridin-2-yl)methylamine and di(pyridin-2-yl)methylamine
Get PDFBy employing the facially coordinated tris(pyridin-2-yl)methylamine ligand, a route to the synthesis of the complex ([Ru(tpyma)(dppz)(py)]2+ (dppz = dipyrido[3,2-a:2′,3′-c]phenazine) py = pyridine) is reported. Detailed NMR spectroscopy and crystallographic studies reveal that tpyma coordinates to the metal ion through a tridentate (2py, NH2), not (3py), binding motif. Subsequently, the synthesis of an analogous complex containing di(pyridin-2-yl)methylamine, dipyma, confirmed that it too is a (2py, NH2) tripodal ligand. In non-aqueous solvents both complexes display emission from the RuII → (dppz)-based 3MLCT excited state. Computational studies reveal that the metal centre of the dipyma complex is more polar than its tpyma analogue and also provide insights into why tpyma does not coordinate to the RuII(dppz) through a (3py) motif. In aqueous solutions the complexes bind to duplex DNA through intercalation with affinities that are entirely comparable to [Ru(bpy)2(dppz)]2+. However, while the tpyma complex displays a DNA light switch effect, the DNA-induced increase in emission from the dipyma complex is significantly lower. This effect is attributed to the different electronic and steric properties of tpyma and dipyma
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! 👇