10784 research outputs found
Sort by
Multi-Domain Adversarial Variational Bayesian Inference for Domain Generalization
No full textPublisher Copyright: © 1991-2012 IEEE.Domain generalization aims to learn common knowledge from multiple observed source domains and transfer it to unseen target domains, e.g. the object recognition in varieties of visual environments. Traditional domain generalization methods aim to learn the feature representation of the raw data with its distribution invariant across domains. This relies on the assumption that the two posterior distributions (the distributions of the label given the feature distribution and given the raw data) are stable in different domains. However, this does not always hold in many practical situations. In this paper, we relax the above assumption by permitting the posterior distribution of the label given the raw data changes in difference domains, and thus focuses on a more realistic learning problem that infers the conditional domain-invariant feature representation. Specifically, a multi-domain adversarial variational Bayesian inference approach is proposed to minimize the inter-domain discrepancy of the conditional distributions of the feature given the label. Besides, it is imposed by the constraints from the adversarial learning and feedback mechanism to enhance the condition invariant feature representation. The extensive experiments on two datasets demonstrate the effectiveness of our approach, as well as the state-of-the-art performance comparing with thirteen methods.Peer reviewe
Corrosion Behavior of Additively Manufactured Steels: A Comprehensive Review
Get PDFPublisher Copyright: © 2025 Wiley-VCH GmbH.Additive manufacturing (AM) is transforming the production of steel components, offering unique advantages such as design freedom and the ability to create complex geometries. This review examines the corrosion behavior of various steel types, including austenitic stainless steels (SS), martensitic SS, duplex SS, low-alloy steels, and maraging steels, produced through AM technologies. In addition, the topic of material hybridization through AM is addressed, which allows for the optimization of the properties of the base materials. While AM often generates finer grain structures, particularly in SS, which enhances corrosion resistance, it can also lead to undesirable phases, precipitates, or defects like porosity that degrade performance. Controlling AM process parameters is crucial to achieving the desired microstructure and optimizing corrosion resistance. The review highlights current knowledge, identifies challenges, and underscores the importance of standardized testing methodologies to enable better cross-study comparisons and guide future advancements in corrosion-resistant AM steels.Peer reviewe
Conversational interfaces, tehcnolanguages and technoinequalities
Get PDFPublisher Copyright: © 2025 Universitat Jaume I. All rights reserved.Conversational interfaces (CIs) enabled by artificial intelligence (AI) technologies promise to reconfigure our relationship with computers through human speech and language as a more natural form of human-machine communication. Using the philosophy of technopersons, the text explores how CIs are promoted by digital platforms to advance a reconfiguration of the social domain upon the possibilities presented by AI technologies for the expansion of their power. This reconfiguration also involves changes into human language and the associated development of technoinequalities. In other words, inequalities reinforced, facilitated and amplified by technology. Four areas of technoinequality associated with the development of CI are identified in terms of linguistic diversity, gender equality, labour precarity and exploitation, and environmental sustainability.Peer reviewe
Heat pump integration for waste heat recovery from a 20 MWe green hydrogen plant to increase global efficiency
Get PDFPublisher Copyright: © 2025 Hydrogen Energy Publications LLCThis paper is focused on the technical and economic analysis of waste heat availability in a 20 MWe green hydrogen (H2) production plant, considering not only waste heat from the polymer electrolyte membrane (PEM) electrolyser cooling circuit, but also from hydrogen and oxygen compression stages. The main objective is to study the behavior of this waste heat along 10 years, until electrolyser End of Life (EoL), representing a maximum of 37.4 % of the total energy consumed by the plant, at an average temperature of 56.5 °C. On the other hand, waste heat upgrade is assessed as a key factor to increase global plant efficiency and so, a strong source to reduce current hydrogen price. In this case, hot water generation through heat pump technology for a 90 °C district heating network is analyzed to define equivalent CO2 emission removal and economic savings, which make the investment technical and economically feasible. At electrolyser EoL, global plant efficiency increases from a present value of 56.5 % up to 90.1 % if heat pump upgrading is considered.Peer reviewe
A machine learning approach for the efficient estimation of ground-level air temperature in urban areas
Get PDFPublisher Copyright: © 2025 Elsevier B.V.The increasingly populated cities of the 21st Century face the challenge of being sustainable and resilient spaces for their inhabitants. However, climate change, among other problems, makes these objectives difficult to achieve. The Urban Heat Island phenomenon that occurs in cities, increasing their thermal stress, is one of the stumbling blocks to achieve a more sustainable city. The ability to estimate temperatures with a high degree of accuracy allows for the identification of the highest priority areas in cities where urban improvements need to be made to reduce thermal discomfort. In this work we posit that image-to-image deep neural networks (DNNs) can effectively correlate spatial and meteorological variables of an urban area with street-level air temperature. To this end, we introduce a novel DNN-based model leveraging a U-Net architecture to tackle this modeling task. We evaluate the proposed model through experiments in a use case focused on the city of Bilbao, Spain. Our method achieves regression performance metrics comparable to those of the numerical model it was trained against, with mean absolute error values below 2°C and a Pearson correlation close to 1. Additionally, it demonstrates strong regression performance against true temperature values recorded by on-site weather stations, enhancing the precision of estimates produced by numerical models. These results confirm that DNNs offer a fast and computationally efficient alternative for the data-driven estimation of ground-level air temperature.Peer reviewe
Decision Support System (DSS) for Manufacturing Engineering of Cans Rolling
No full textPublisher Copyright: © The Author(s) 2025.Decision Support Systems (DSS) can help factory workers in the decision-making step of multiple tasks. In digital factories, these systems make use of data towards a human-centered manufacturing. Rolling of large and thick plates into cans is a common practice in the metal forming industry to fabricate pipes or tanks. The process is adjusted by trial and error with a high level of operator intervention. Furthermore, only a small number of cans are identical. The objective of this work is to prescribe, by means of a DSS, the process parameters to be applied by the operator in the machine to optimize the can fabrication. The development of the DSS involved several steps, including firstly signal preprocessing and classification and then data extraction, aggregation, and regression in a multi-stage prediction framework. A significant use of domain knowledge for a data-centric solution contributes to the quality of the recommendations and the ability to organize and transfer know-how among operators.Peer reviewe
Urban Data Governance: an Interoperability-Based Approach for Monitoring Natural Threats at Different Geographic Scales, Through Smart City Platforms
No full textPublisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.Monitoring, preventing and managing the impacts induced by extreme natural events requires the use of multiple and different Information Communication Technology (ICT) tools and technologies capable of collecting and processing data from various sources, and supporting public stakeholders in the planning and implementation of prompt actions. In this perspective, the availability of platforms able to harmonize, integrate and manage heterogeneous data and to create new knowledge, can constitute a valuable support. This article presents some preliminary results from the EU-funded MULTICLIMACT project, part of which is defining a reference model for customizing and adopting the ENEA Smart City Platform to inform on the severity and extent of possible impacts induced by natural threats and on possible resilience strategies. Towards that the Smart City Platform enables the interoperability between heterogeneous digital solutions monitoring natural threats at different geographic scales.Peer reviewe
A decade review on hardwood composites and their research developments
No full textPublisher Copyright: © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.Composite materials, such as engineered wood products (EWPs), are favored for their stability, durability, and uniform mechanical properties, making them a potential way to valorize hardwood species. Therefore, this article aims to present a decade review of hardwood composites and their developments. The methodology of this review was divided into three main parts. First, a search for peer-reviewed articles on EWPs published between 2012 and 2023 was conducted using the Web of Science (WOS) database. Second, the data extracted from the WOS database were analyzed using the Virtual Operating System (VOS) viewer and a bibliometric approach. Third, selected peer-reviewed articles were systematically reviewed and included in this study, and their main findings were presented. The reviewed documents showed that hardwood composites are favored for their stability, durability, and uniform mechanical properties, which make them attractive for various engineering applications. Some key findings from the review include the potential of Yellow Birch to enhance wood-plastic composites by improving flexural strength and reducing flammability. Eucalyptus nanofibers have shown promise in enhancing the mechanical performance of composite mixtures. Hybrid poplar has been identified for its suitability in cross-laminated timber (CLT) products as they meet and exceed the shears and bending strength required by ANSI/APA PRG-320. Moreover, I-joists made from hardwood residues exhibit comparable mechanical performance to commercial counterparts. The review underscores the need for continued research and development to foster wider adoption of these valuable materials.Peer reviewe
WoodAD: A New Dataset and a Comparison of Deep Learning Approaches for Wood Anomaly Detection
No full textPublisher Copyright: © 2025 John Wiley & Sons Ltd.Anomaly detection is a crucial task in computer vision, with applications ranging from quality control to security monitoring, among many others. Recent technological advancements have enabled near-perfect solutions on benchmark datasets like MVTec, raising the need for novel datasets that pose new challenges for this modelling task. This work presents a novel Wood Anomaly Detection (WoodAD) dataset, which includes defects in wooden pieces that result in challenges for the most advanced techniques applied to other established datasets. This article evaluates such challenges posed by WoodAD with one-class and few-shot supervised learning approaches. Our experiments herein reveal that EfficientAD, a state-of-the-art method previously excelling on the MVTec dataset, outperforms all other one-class learning approaches. Nevertheless, there is room for improvement, as EfficientAD achieves a 0.535 pixel/segmentation average precision (AP) over the complete test set. UNet, a well-known pixel-level classification architecture, leveraged few-shot supervised learning to enhance the pixel AP score, achieving 0.862 pixel/segmentation AP over the entire test set. Our WoodAD dataset represents a valuable contribution to the field of anomaly detection, offering complex image textures and challenging defects. Researchers and practitioners are encouraged to leverage this dataset to push the boundaries of anomaly detection and develop more robust and effective solutions for more complex real-world applications. The WoodAD dataset has been made publicly available in Kaggle (https://www.kaggle.com/datasets/itiresearch/wood-anomaly-detection-one-class-classification).Peer reviewe
Recovery and purification of acetic acid from extremely diluted solutions using a mixed bed ion exchange resin - technical feasibility
Get PDFPublisher Copyright: © 2025 The Royal Society of Chemistry.A downstream process for the recovery and purification of acetic acid (AA) from an extremely diluted solution (100 mg L−1) also containing a mixture of contaminating inorganic salts in the form of bicarbonates, phosphates, sulfates and chlorides (DPM medium) has been developed, showing its technical feasibility. The process involves two successive steps based on the use of a mixed bed ion exchange (IEX) resin. The first step, a demineralization treatment to remove the inorganic anions that could potentially interfere with the recovery and purification of AA, involves a combined treatment of calcium precipitation, acidification with the Amberlite IR-120 resin and treatment with the Amberlite MB20 mixed bed resin. This treatment allows the total removal of phosphate and sulfate (and likely bicarbonate) and 90% removal of chloride, while still retaining 91% of AA in solution. In the second step the demineralized medium is treated again with the Amberlite MB20 mixed bed resin in batch to completely remove AA and chloride remaining in solution and, finally, the anion-loaded resin is step-eluted with a low volume of diluted H2SO4 to selectively elute AA, obtaining a purified (68.5-82.2% recovery yield and 96.9-99.2% purity) and concentrated (>1500 mg L−1) solution of the acid.Peer reviewe