30793 research outputs found
Sort by
Bonding Defect Detection Based on Improved Single Shot MultiBox Detector
Get PDFTo solve the problem of time-consuming and low efficiency in manual defect detection, this paper proposes a bonding defect detection algorithm based on improved Single Shot MultiBox Detector (SSD). DenseNet is used to replace VGG of the SSD algorithm to improve the detection effect of bonding defect. A novel feature fusion network is designed, in which dilated convolution is used to reduce the size of the low-level feature map, and it is fused with the high-level feature map, and then the Convolutional Block Attention Module (CBAM) attention mechanism is used to increase the ability to extract the features. Focal loss is used to control the ratio of positive and negative samples for training and suppress easily separable samples, so that the samples involved in training have better distribution and the model has better detection performance. Then, the defect data set is constructed and a comparison experiment is carried out. The results show that the mAP, Precision, and Recall of the improved SSD network are increased to 75.9 %, 77.3 %, and 75.6 %, respectively, which can better identify bonding defect.This work was supported by the Research Project supported by the Shanxi Scholarship
Council of China under Grant No. 2022-145
Detailed Image Captioning and Hashtag Generation
Get PDFData Availability Statement:
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.This article presents CapFlow, an integrated approach to detailed image captioning and hashtag generation. Based on a thorough performance evaluation, the image captioning model utilizes a fine-tuned vision-language model with Low-Rank Adaptation (LoRA), while the hashtag generation employs the keyword extraction method. We evaluated the state-of-the-art image captioning models using both traditional metrics (BLEU, METEOR, ROUGE-L, and CIDEr) and the specialized CAPTURE metric for detailed captions. The hashtag generation models were assessed using precision, recall, and F1-score. The proposed method demonstrates competitive results against larger models while maintaining efficiency suitable for real-time applications. The image captioning model outperforms the base Florence-2 model and favorably compares with larger models. The KeyBERT implementation for hashtag generation surpasses other keyword extraction methods in both accuracy and speed. This work contributes to the field of AI-assisted content analysis and generation, offering insights into the practical implementation of advanced vision-language models for detailed image understanding and relevant tag generation.This research received no external funding
Reform or transform? A spectrum of stances towards the economic status quo within ‘new economics’ discourses
Get PDFData availability statement:
The authors take responsibility for the integrity of the data and the accuracy of the analysis, although errors in data extracted directly from Scopus and Google may be beyond our control. See Supplementary data (available on figshare) for access to the literature sample data, or contact the corresponding author.Key messages:
• New economics ‘discourse coalitions’ are needed to replace neoliberalism and transcend global crises.
• New economics discourses show variation in the depth of change advocated.
• Most authors align with the SDGs and against neoliberalism. Stances diverge on economic growth.
• There is a general lack of explicit mentions of key aspects of the economic status quo.‘New economics’ discourses – comprising diverse approaches advocated as more just and sustainable replacements of dominant neoclassical and neoliberal economic perspectives – have been criticised as insufficiently coherent to form the ‘discourse coalitions’ necessary to enter the mainstream. To date there has been little systematic exploration of the agreement or divergence in new economics discourses. Here, we conduct a qualitative systematised review of new economics literature in the context of the COVID-19 pandemic to analyse stances towards the economic status quo and the depth of change advocated in it, such as fundamental and systemic transformation or more superficial reformist or accepting types of change that mostly maintain current economic systems. We interpreted authors’ stances towards six key status quo themes: capitalism; neoliberalism; GDP-based economic growth; debt-based money; globalisation; and the Sustainable Development Goals (SDGs). In the 525 documents analysed, there was relative consensus that neoliberalism needed transforming, stances towards GDP-based growth substantially diverged (from transformative to reformist/accepting), and stances towards the SDGs were mostly accepting, although the status quo themes tended to be infrequently mentioned overall. Different new economics approaches were associated with diverging stances. We suggest that alignment against neoliberalism and towards the SDGs may provide strategic coalescing points for new economics. Because stances towards core problematised aspects of mainstream economics were often not articulated, we encourage new economics scholars and practitioners to remain explicit, aware and reflexive with regard to the economic status quo, as well as strategic in their approach to seeking economic transformation.This work is part of the Global Assessment for a New Economics (GANE) project supported by the Leverhulme Trust-funded Leverhulme Centre for Anthropocene Biodiversity under Grant RC-2018-021, the Laudes Foundation, the University of York and Ecologos Research Ltd
Design of Scalable Population of Reinforcement Learning Agents for Autonomous 5G Radio Link Control
Get PDFThis research demonstrates how MATLAB's Reinforcement Learning Markov Decision Process (MDP) Example Model can be used to design Radio Link Control MDP Reinforcement Learning (RL) agent. Since the number of agents in MATLAB's RL toolbox is not scalable beyond one agent, then an agent scalability scheme is required to design RL agents in MATLAB's RL toolbox and then realize multiple lightweight simultaneously operable Python instances of it for each of the multiple user equipment UE in a network.The authors gratefully acknowledge support of EU Horizon 2020 Research Project 6G BRAINS (Bringing Reinforcement learning Into Radio Light Network for Massive Connections)
Explainable Semantic Federated Learning Enabled Industrial Edge Network for Fire Surveillance
Get PDFIn fire surveillance, Industrial Internet of Things (IIoT) devices require transmitting large monitoring data frequently, which leads to huge consumption of spectrum resources. Hence, we propose an Industrial Edge Semantic Network to allow IIoT devices to send warnings through Semantic communication (SC). Thus, we should consider 1) data privacy and security; 2) SC model adaptation for heterogeneous devices; 3) explainability of semantics. Therefore, first, we present an eXplainable Semantic Federated Learning (XSFL) to train the SC model, thus ensuring data privacy and security. Then, we present an adaptive client training strategy to provide a specific SC model for each device according to its Fisher information matrix, thus overcoming the heterogeneity. Next, an Explainable SC mechanism is designed, which introduces a leakyReLU-based activation mapping to explain the relationship between the extracted semantics and monitoring data. Finally, simulation results demonstrate the effectiveness of XSFL.10.13039/501100001809-National Natural Science Foundation of China (Grant Number: 41904127, 41604117 and 62132004);
10.13039/501100004735-Natural Science Foundation of Hunan Province (Grant Number: 2024JJ5270);
Open Project of Xiangjiang Laboratory (Grant Number: 22XJ03011);
Scientific Research Fund of Hunan Provincial Education Department (Grant Number: 22B0663);
Changsha Natural Science Foundation (Grant Number: kq2402098 and kq2402162)
Resilience and sustainability assessment of a prestressed concrete viaduct
Get PDFConference paper presented at 12th International Conference on Bridge Maintenance, Safety and Management (IABMAS 2024, Copenhagen, Denmark, 24-28 June 2024).The paper explores Greece’s Polyfytos viaduct, the nation’s second-longest bridge spanning 1,372 meters, renowned also for providing access to key power plants in South-East Europe. A resilience analysis of different retrofitting scenarios was conducted, including both transitability and structural capacity aspects, employing visual inspections, digital data collection, and advanced modeling techniques. The viaduct, conceptualized by Prof Riccardo Morandi and built in 1972-1975, displays typical degradation seen in prestressed and reinforced concrete bridges. Unlike previous research focused solely on retrofitting with LCA and LCC assessments, this study integrates resilience assessment, marking a pioneering holistic approach to viaducts refurbishment.The research has received funding from the (i) ERC Grant ID: 101007595 of the project ADDOPTML, MSCA RISE 2020; (ii) EU HORIZON-MSCA-2021-SE-01 Grant No: 101086413, ReCharged. Stergios Mitoulis and Sotirios Argyroudis received funding from the UK Research and Innovation (UKRI) under the Horizon Europe funding guarantee (agreements No. EP/Y003586/1, EP/X037665/1), for the project ‘ReCharged - Climate-aware Resilience for Sustainable Critical and interdependent Infrastructure Systems enhanced by emerging Digital Technologies’ (Grant No: 101086413)
Resilience Analysis of Different Retrofitting Solutions for a Prestressed Concrete Viaduct
Get PDFConference paper presented at the 2nd Fabre Conference – Existing bridges, viaducts and tunnels: research, innovation and applications (FABRE24), Genoa, Italy, 12-15 February 2024.This article introduces a resilience analysis conducted on Greece’s prestressed Polyfytos viaduct. As the nation’s second longest bridge, spanning 1,372 meters, it was conceptualized by Prof Riccardo Morandi and built between 1972 and 1975, marking it as an iconic structure at 48 years old. Notably, the bridge has strong interdependencies with key power plants, dams and solar pars in the region. Evaluating both risk and resilience, the study employed visual inspections and digital data collection methods. These methods involved a digital twin, offering current asset geometry and a dynamic model for advanced simulations; satellite imagery for ongoing updates on the structure’s deformations and geometry; and advanced numerical modeling aimed at interpreting current deflections via back analysis. The bridge shows signs of degradation commonly found in reinforced concrete (RC) and prestressed RC (PRC) bridges, specifically concerning corroded tendons and concrete bonding. Prior research focused on evaluating various retrofitting approaches and their lifecycle impacts, whereas this study integrates the resilience assessment of such retrofit solutions. This contribution represents a new step in the direction of a holistic approach to identifying the appropriate retrofit of an existing viaduct aiming to inform decision-making about the benefits of different restoration investments.The research has received funding from the (i) ERC Grant ID: 101007595 of the project ADDOPTML, MSCA RISE 2020; (ii) EU HORIZON-MSCA-2021-SE-01 Grant No: 101086413, ReCharged. Stergios Mitoulis and Sotirios Argyroudis received funding by the UK Research and Innovation (UKRI) under the UK government’s Horizon Europe funding guarantee for the European Union HORIZON-MSCA-2021-SE-01, grant agreement No: 101086413, ReCharged
Impact of Lewy bodies disease on visual skills and memory abilities: from prodromal stages to dementia
Get PDFDementia with Lewy bodies (DLB) and its prodromal presentation with mild cognitive impairment is characterized by prominent deficits in attention/executive domains and in visual processing abilities with relative sparing of memory. Neuropsychological research is continuously refining the tools to define more in detail the patterns of relatively preserved and impaired cognitive abilities that help differential diagnosis between DLB and Alzheimer disease (AD). This review summarizes the main studies exploring specific cognitive tasks investigating different visual processing abilities and verbal memory that better differentiate DLB from AD. The findings provide evidence that substantial impairments in visual-spatial and visual-constructional abilities and relatively better performance on memory tasks that depend on hippocampal function characterize the prodromal stage of DLB. The ability to detect early indicators of prodromal DLB through clinical and cognitive assessments is the first step to guide instrumental diagnostic work-ups and provide the opportunity for early intervention.Open Access funding provided by Università degli Studi di Padova | University of Padua, Open Science Committee. Acknowledgements: CB is supported by a liberal donation by Fondazione Leo Pavan. AV and MM are supported by funding obtained under the National Recovery and Resilience Plan (NRRP), Mission 4 Component 2 Investment 1.3 - Call for tender No. 341 of 15/03/2022 of Italian Ministry of University and Research funded by the European Union – NextGenerationEU, Project code PE0000006, Concession Decree No. 1553 of 11/10/2022 adopted by the Italian Ministry of University and Research, CUP D93C22000930002, “A multiscale integrated approach to the study of the nervous system in health and disease” (MNESYS). MM acknowledges the support by funding from the Italian Ministry of Health (#GR-2019-12369242)
Rapid post-disaster infrastructure damage characterisation using remote sensing and deep learning technologies: A tiered approach
Get PDFData availability:
No data was used for the research described in the article.Critical infrastructure is vital for connectivity and economic growth but faces systemic threats from human-induced damage, climate change and natural disasters. Rapid, multi-scale damage assessments are essential, yet integrated, automated methodologies remain underdeveloped. This paper presents a multi-scale tiered approach, which addresses this gap, by demonstrating how automated damage characterisation can be achieved using digital technologies. The methodology is then applied and validated through a case study in Ukraine involving 17 bridges damaged by targeted human interventions. Technology is deployed across regional to component scales, integrating assessments using Sentinel-1 SAR images, crowdsourced data, and high-resolution images for deep learning to enable automatic damage detection and characterisation. The interferometric coherence difference and semantic segmentation of images are utilised in a tiered multi-scale approach to enhance the reliability of damage characterisation at various scales. This integrated methodology automates and accelerates decision-making, facilitating more efficient restoration and adaptation efforts and ultimately enhancing infrastructure resilience.The first author would like to acknowledge the financial supports from British Academy for this research (Award Reference: RaR\100770).
Dr. Stergios-Aristoteles Mitoulis and Dr. Sotirios Argyroudis received funding by the UK Research and Innovation (UKRI) under the UK government's Horizon Europe funding guarantee [Ref: EP/Y003586/1, EP/X037665/1]. This is the funding guarantee for the European Union HORIZON-MSCA-2021-SE-01 [grant agreement No: 101086413] ReCharged - Climate-aware Resilience for Sustainable Critical and interdependent Infrastructure Systems enhanced by emerging Digital Technologies
Innovative machine learning approaches for indoor air temperature forecasting in smart infrastructure
Get PDFData availability:
All data are fully available without restriction. They may be found at: https://data.mendeley.com/datasets/fj23wz552c/1.Caro, Rosana (2024), “MunicipalitiesAndalusiaSpain2019_SimulatedMonitored”, Mendeley Data, V1, doi: 10.17632/fj23wz552c.1.Efficient energy management and maintaining an optimal indoor climate in buildings are critical tasks in today’s world. This paper presents an innovative approach to surrogate modeling for predicting indoor air temperature (IAT) in buildings, leveraging advanced machine learning techniques. At the core of this study is the application of Long Short-Term Memory (LSTM) networks for time-series modeling, which significantly enhances the capture of temporal dependencies in temperature predictions. The proposed LSTM with RWCV (Rolling Window Cross-Validation) offers significant advantages over a usual LSTM in time-series tasks, particularly due to its ability to adapt to new data trends through the rolling window mechanism. It provides more robust and generalizable forecasts in dynamic environments, prevents overfitting through dropout and cross-validation, and improves model evaluation with temporal integrity. In contrast, traditional LSTM models are better suited for static, non-evolving datasets and may not handle dynamic time-series data effectively. To rigorously assess model performance, a comprehensive evaluation framework is developed, incorporating metrics such as mean square error (MSE) and the coefficient of determination (R²). Additionally, a novel cumulative error analysis method is introduced enabling real-time monitoring and model adjustment to maintain predictive accuracy over time. Test results demonstrate that model losses on the test dataset are only marginally higher than those on the training dataset, indicating robust generalization capabilities. Loss values range from 0.0004709 to 0.02819861, depending on building operating conditions. A comparative analysis reveals that Adaboost and Gradient Boosting models outperform linear regression, highlighting their potential for achieving energy-efficient and comfortable indoor climate management in buildings. The findings underscore the efficacy of the proposed approach for IAT prediction and point towards further research possibilities in dataset expansion and model optimization to enhance building climate management and energy conservation.This research is funded by the HORIZON EUROPE project ZEBAI: Innovative methodologies for the design of Zero-Emission and cost-effective Buildings enhanced by Artificial Intelligence (Grant agreement ID: 101138678)
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! 👇