10784 research outputs found
Sort by
Taxonomic hierarchical loss function for enhanced crop and weed phenotyping in multi-task semantic segmentation
Full text linkPublisher Copyright: © 2025 The Author(s)Herbicide research and development necessitate specific trials to monitor the effects of various herbicide formulations, quantities, and protocols on different plant species and growth stages. These trials are necessary to ensure the safety and efficacy of the developed products. Currently, these tests are conducted manually and assessed visually, making the process time-consuming and labor-intensive. Developing a computer model to characterize species, damage, and growth stages is challenging due to the fine-grained differences between species and damage, significant intra-class variability, and difficulties in manual annotations. Additionally, manually annotated datasets for semantic segmentation are often imperfect. The presence of non-target or unknown species, where only the genus or family is known, complicates the management and scalability of these datasets. In this work, we propose a new hierarchical loss function, suitable for semantic segmentation tasks, capable to take advantage for the hierarchical taxonomy relationships between species, plant damages and other relationships and thus, reduce the need for annotated data. The proposed loss function support datasets with varying granularity and annotation heterogeneity, including for partial annotations at the pixel level. We validated this loss function using a multi-task semantic segmentation neural network to simultaneously detect plant species and quantify the damage of each species. The proposed hierarchical loss function improves model performance, increasing the F1-Score for species detection from 0.41 to 0.52, for damage detection from 0.23 to 0.28. This enhancement forces the model to learn richer hierarchical representations, enabling the support of heterogeneous and partially annotated scalable datasets, which are common in real-world AI applications.Peer reviewe
Novel Binders for Aqueous Electrode Processing of Electrochemical Capacitors
No full textPublisher Copyright: © 2024 The Authors. ChemSusChem published by Wiley-VCH GmbH.This work studies the use of epoxy and polyurethane formulations as binders for the aqueous processing of activated carbon (AC) electrodes used as positive and negative electrodes in Electrochemical Double Layer Capacitors (EDLCs). The use of amine and carbodiimide as crosslinkers is also evaluated. The mechanical properties of those different binders have been investigated, looking towards aqueous processable and flexible electrodes. Microstructural analysis of the fabricated AC electrodes has been carried out to understand the pore-blocking effect exhibited by certain polymers. Furthermore, electrochemical characterization of all the systems has been performed by cyclic voltammetry, electrochemical impedance spectroscopy, and constant current charge/discharge measurements at different current densities. The obtained results show that polyurethane (PU) outperforms in terms of energy and power density the carboxymethyl cellulose:styrene butadiene rubber (CMC : SBR) reference system. Moreover, the studied polyurethanes maintain close to 100 % of their initial capacitance after 2500 cycles under a current density of 5 A g−1 and a discharge time of 20 s.Peer reviewe
Improving the power production of Mutriku Wave Power Plant throughout tuning the actual generator and damping valve limits
Full text linkPublisher Copyright: © 2024 The Author(s)The Mutriku Wave Power Plant (MWPP) has successfully operated since 2011, demonstrating remarkable operational stability with minimal issues of OWC technology in breakwaters. So far, the design, selection, and operation ranges of the generators at MWPP prioritized reliability. However, once the viability of the technology has been demonstrated, it is important to explore ways to increase energy production with the current design of the technology. The objective of the present work is to improve the power production in MWPP throughout the Power take-off (PTO), increasing the operational limits of the generator and of the damping valve, which is located between the air chamber and the turbine. To this aim, new sensors have been installed in the plant, such as an inlet pressure sensor, to characterise the behaviour of the PTO. Investigations into the power produced and the plant availability have concluded that modifying control laws of the current configuration could help to increase power production under generator thermal operation ranges. The numerical results presented in the paper demonstrate that the potential benefit to the overall energy production of the MWPP could be significant, leading to greater advancements in the feasibility of wave energy technologies. The analysis of changes and improvements into the MWPP control strategies will provide further guidance into the development of novel wave energy control systems and components for future testing.Peer reviewe
Experimental characterization of the mechanical and functional performance of innovative ultra-low carbon sandwich panels and envelope systems for buildings
Full text linkPublisher Copyright: © 2025 The Author(s)The use of low-carbon materials is key for decarbonizing the construction sector, but their innovation potential is still relatively unexplored. The EU research project InCSEB aimed at the development of five ultra-low carbon steel building envelope systems and sandwich panels for cladding and roofing applications. These façade systems and sandwich panels incorporate the innovative use of wood fiber, a renewable and bio-sourced insulation material, while achieving a high level of mechanical and functional performances and ensuring compliance with other requirements such as sustainability criteria. This paper provides an overview of the results of the experimental study that has been carried out to assess and validate the functional and mechanical performances of the proposed innovative insulation sandwich panels and façade systems, aiming to achieve a technology readiness level of TRL 7 by the end of the project. The experimental study shows very promising results, fulfilling market expectations regarding mechanical, seismic, thermal, environmental, and regulatory requirements.Peer reviewe
On the caveats of AI autophagy
No full textPublisher Copyright: © Springer Nature Limited 2025.Generative artificial intelligence (AI) technologies and large models are producing realistic outputs across various domains, such as images, text, speech and music. Creating these advanced generative models requires significant resources, particularly large and high-quality datasets. To minimize training expenses, many algorithm developers use data created by the models themselves as a cost-effective training solution. However, not all synthetic data effectively improve model performance, necessitating a strategic balance in the use of real versus synthetic data to optimize outcomes. Currently, the previously well-controlled integration of real and synthetic data is becoming uncontrollable. The widespread and unregulated dissemination of synthetic data online leads to the contamination of datasets traditionally compiled through web scraping, now mixed with unlabelled synthetic data. This trend, known as the AI autophagy phenomenon, suggests a future where generative AI systems may increasingly consume their own outputs without discernment, raising concerns about model performance, reliability and ethical implications. What will happen if generative AI continuously consumes itself without discernment? What measures can we take to mitigate the potential adverse effects? To address these research questions, this Perspective examines the existing literature, delving into the consequences of AI autophagy, analysing the associated risks and exploring strategies to mitigate its impact. Our aim is to provide a comprehensive perspective on this phenomenon advocating for a balanced approach that promotes the sustainable development of generative AI technologies in the era of large models.Peer reviewe
Efficient bridge damage detection using a lightweight attention-based modeling framework
Full text linkPublisher Copyright: © 2025 The Author(s). Computer-Aided Civil and Infrastructure Engineering published by Wiley Periodicals LLC on behalf of Editor.Currently, real-time assessment of surface damage to bridges is crucial for ensuring infrastructure safety. Unfortunately, existing methods often present a challenge: overly complex computational models are incompatible with systems that have limited resources, while lightweight models struggle to achieve sufficient detection accuracy. This task is further complicated by the diverse nature of bridge damages, such as cracks, exposed reinforcement, and efflorescence, as well as the challenges of data acquisition under varied conditions from sources like unmanned aerial vehicles and specialized datasets. This work presents an efficient framework developed to improve such applications. The Lightweight Feature Enhancement and Triplet Attention Network for Bridge Damage Detection includes: (1) a multi-scale feature learning module, (2) a slim-neck-based optimized feature pyramid integration module, and (3) a triplet attention-based damage detector module; (1) extracts multi-scale representations of bridge surface features, (2) enhances multi-scale feature integration for lightweight computation, while maintaining accuracy, and (3) optimizes the framework with a three-branch structure for cross-latitude interaction, reducing the importance of irrelevant features. Extensive experiments on the MCDS and CODEBRIM datasets demonstrated its advantages: a (Formula presented.) increase in Mean Average Precision, a (Formula presented.) computational load reduction, and a 45 frames per second real-time performance. The model's computational complexity scales linearly with the input instances processed per unit time during inference.Peer reviewe
Crossover of critical behavior in dynamic phase transitions of multilayer Ising model systems
No full textPublisher Copyright: ©2025 American Physical Society.We investigate the crossover of critical behavior for the dynamic phase transition (DPT) in ferromagnetic thin films using Monte Carlo simulations of the kinetic Ising model, focusing on the scaling behavior of the dynamic order parameter under a time-dependent external magnetic field. Specifically, we study the transition of the critical behavior of such multilayer film systems from two-dimensional (2D) to three-dimensional (3D) as a function of the film thickness and the distance to the critical point, which enables dimensional crossover observations. Our results indicate that the effective critical exponents exhibit a clear transition in their scaling behavior, with thinner films showing 2D-like characteristics and thicker films displaying 3D-like behavior, for both the DPT and the thermodynamic phase transitions (TPT). Quantitatively, the crossover from 2D to 3D behavior occurs at larger film thicknesses for the DPT compared to the TPT, suggesting that DPT and TPT are governed by distinctly different length scales and underlying surface effects. These findings are in agreement with experimental observations in ultrathin Co films, where dynamic and thermodynamic critical exponents were found to differ. Therefore, our study provides an in-depth explanation for critical phenomena in thin-film ferromagnets driven by a time-dependent magnetic field. By comparing the dimensional crossover properties of both TPT and DPT, we present a comprehensive understanding of how thin-film geometry and surface effects influence the scaling laws and critical behavior in nonequilibrium systems.Peer reviewe
Low-cost optical fiber multimode interference biosensor based on a glucose sensitive Glucose-Oxidase enzyme thin-film
No full textPublisher Copyright: © 2024In this research we report a contribution for the development of low-cost fiber optical biosensors fabricated by the Single Mode-Multi Mode-Single Mode configuration applied for the glucose monitoring considering clinical concentrations ranges in aqueous analytes. Designed devices are evaluated using health standard detection ranges, such as healthy, pre-diabetic, and diabetic stages operating at the visible spectral region. The sensing regions has been prepared by the etching technique in order to improve the interaction between the evanescent wave with the surrounding medium followed by functionalization of enzyme oxidase glucose via the electrostatic self-assembly using by Poly(allylamine hydrochloride) as an immobilizer matrix. The increase of bilayers number over the sensor surface permits us to demonstrate the enhancement of sensitivity and limit of detection. Experimental results permitted the glucose characterization in the range from 0.3 to 2.4 mg mL-1 obtaining a response time of 9 s and a sensitivity of 1.8 nm/(mg mL-1) allowing to detect hypoglycemia and diabetes stages according to the World Health Organization standards.Peer reviewe
Source-dependent absorption Ångström exponent in the Los Angeles Basin: Multi-time resolution factor analyses of ambient PM2.5 and aerosol optical absorption
No full textPublisher Copyright: © 2024 The AuthorsAdvanced receptor models can leverage the information derived from optical and chemical variables as input by a variety of instruments at different time resolutions to extract the source specific absorption Ångström exponent (AAE) from aerosol absorption. The multilinear engine (ME-2), a Positive Matrix Factorization (PMF) solver, serves as a proficient tool for performing such analyses, thereby overcoming the constraints imposed by the assumptions in current optical source apportionment methods such as the Aethalometer approach since the use of a-priori AAE values introduces additional uncertainty into the results of optical methods. Comprehensive PM2.5 chemical speciation datasets, and aerosol absorption coefficients (babs, λ) at seven wavelengths measured by an Aethalometer (AE33), were used in multi-time source apportionment (MT-PMF). The study focused on two locations in the Los Angeles (LA) Basin: Central LA (CELA, Main St.), an urban area surrounded by major freeways, and Rubidoux (RIVR), a residential area surrounded by local roads. Factor profiles and temporal variations of their contributions were obtained. Additionally, factor displacements (DISP) and profile constraints were applied. Five-factor solutions were obtained at both sites. At CELA, the resolved factors included traffic + crustal matter (traffic+ Cr_M), secondary sulfate + nitrate (SSN), biomass burning (BB), diesel emissions (DIE) and aged sea salt (ASS). Moreover, source-dependent AAE values at CELA were obtained without a-priori assumption, with values of 1.46 for traffic+ Cr_M, 1.45 for DIE and 2.37 for BB. At RIVR, the resolved factors were traffic+ Cr_M (AAE = 1.24), particulate sulfate, particulate nitrate, BB (AAE = 3.00) and aged sea salt. PM2.5 composition differed at both locations. SSN accounted for the largest fraction of the ambient PM2.5 mass concentration, their sum at the CELA site averaged 40 % of the PM2.5 mass while the same species represented 77 % at RIVR.Peer reviewe
