Applied Sciences

19 pages, 773 KiB

Open AccessArticle

Explainable AI-Enhanced Human Activity Recognition for Human–Robot Collaboration in Agriculture

by Lefteris Benos, Dimitrios Tsaopoulos, Aristotelis C. Tagarakis, Dimitrios Kateris, Patrizia Busato and Dionysis Bochtis

Appl. Sci. 2025, 15(2), 650; https://doi.org/10.3390/app15020650 (registering DOI) - 10 Jan 2025

Abstract

This study addresses a critical gap in human activity recognition (HAR) research by enhancing both the explainability and efficiency of activity classification in collaborative human–robot systems, particularly in agricultural environments. While traditional HAR models often prioritize improving overall classification accuracy, they typically lack [...] Read more.

This study addresses a critical gap in human activity recognition (HAR) research by enhancing both the explainability and efficiency of activity classification in collaborative human–robot systems, particularly in agricultural environments. While traditional HAR models often prioritize improving overall classification accuracy, they typically lack transparency in how sensor data contribute to decision-making. To fill this gap, this study integrates explainable artificial intelligence, specifically SHapley Additive exPlanations (SHAP), thus enhancing the interpretability of the model. Data were collected from 20 participants who wore five inertial measurement units (IMUs) at various body positions while performing material handling tasks involving an unmanned ground vehicle in a field collaborative harvesting scenario. The results highlight the central role of torso-mounted sensors, particularly in the lumbar region, cervix, and chest, in capturing core movements, while wrist sensors provided useful complementary information, especially for load-related activities. The XGBoost-based model, selected mainly for allowing an in-depth analysis of feature contributions by considerably reducing the complexity of calculations, demonstrated strong performance in HAR. The findings indicate that future research should focus on enlarging the dataset, investigating the use of additional sensors and sensor placements, and performing real-world trials to enhance the model’s generalizability and adaptability for practical agricultural applications. Full article

(This article belongs to the Special Issue Explainable Artificial Intelligence for Visualization in Human Computer Interactions)

24 pages, 521 KiB

Open AccessArticle

A Classification Model Based on Interval Rule Inference Network with Interpretability

by Yunxia Zhang, Yiming Zhong, Xiaochang Wu and Jing Bai

Appl. Sci. 2025, 15(2), 649; https://doi.org/10.3390/app15020649 (registering DOI) - 10 Jan 2025

Abstract

Interpretability requirements, complex uncertain data processing, and limited training data are characteristics of classification in some real industry applications. The interval belief rule base (IBRB) can deal with various types of uncertainty and provides high interpretability. However, there is a large number of [...] Read more.

Interpretability requirements, complex uncertain data processing, and limited training data are characteristics of classification in some real industry applications. The interval belief rule base (IBRB) can deal with various types of uncertainty and provides high interpretability. However, there is a large number of parameters in IBRB, which makes it difficult for experts to accurately set them manually, limiting its application scope. To address this issue, this paper proposes an interval rule inference network (IRIN) with interpretability for classification models to automatically generate IBRB through integrating the ideas of the IBRB and the neural network. Firstly, hybrid data with different types are transformed into an interval belief distribution for automatic generation processing. Secondly, the interval evidence reasoning method is utilized as the inference engine to transfer information ensuring the process’s interpretability. Finally, a reasonable IBRB is generated automatically by updating the parameters by employing the learning engine in the neural network. Moreover, the differentiability of the interval evidence reasoning method in the IRIN is proved as a theoretical foundation of the IRIN, and an interpretability analysis of the IRIN’s structures is discussed. Experimental results demonstrate that the proposed method possesses high interpretability, enhancing the reliability of classification and maintaining the accuracy. Its application in an actual engineering case illustrates that it is particularly suitable for engineering problems where the explanation of results is a critical requirement. Full article

16 pages, 3784 KiB

Open AccessArticle

The Effect of Variable Ratios of Na₂O/K₂O Oxides in Glazes Containing BaO, ZnO, and ZrO₂: Structural Analysis, Characteristic Temperatures, and Surface Properties

by Janusz Partyka, Dawid Kozień and Katarzyna Pasiut

Appl. Sci. 2025, 15(2), 648; https://doi.org/10.3390/app15020648 - 10 Jan 2025

Abstract

In this paper, the glazes of a multicomponent system (SiO₂-Al₂O₃-CaO-MgO-Na₂O-K₂O-BaO-ZnO-ZrO₂) were examined. This work focuses on five glazes and the difference between them as the molar ratio of the alkali oxides [...] Read more.

In this paper, the glazes of a multicomponent system (SiO₂-Al₂O₃-CaO-MgO-Na₂O-K₂O-BaO-ZnO-ZrO₂) were examined. This work focuses on five glazes and the difference between them as the molar ratio of the alkali oxides Na₂O/K₂O. Analysis of fired glazes focused on changes in phase composition (qualitative and quantitative) microstructure performed during observations made by scanning electron microscopy (SEM). The changing molar ratios were also studied in the structure analysis based on the result of data obtained by middle infrared (MIR) and Raman spectroscopy. The characteristic temperatures of the analyzed glazes were also designated using high-stage microscopy. Surface properties such as the color and roughness of the fired glazes were measured by means of a spectrometer and confocal microscopy as well. The amount and type of crystalline phases with the molar ratio of alkali oxides in the analyzed glazes were changed. In the glazes, the crystalline phase of a solid solution of plagioclase was obtained. The results obtained indicate that glazes with a predominant potassium oxide are characterized by lower characteristic temperatures and greater surface smoothness. Structure analysis indicates a different role for the five-molar ratio of Na₂O/K₂O. Full article

(This article belongs to the Special Issue Novel Ceramic Materials: Processes, Properties and Applications)

57 pages, 895 KiB

Open AccessReview

Privacy Auditing in Differential Private Machine Learning: The Current Trends

by Ivars Namatevs, Kaspars Sudars, Arturs Nikulins and Kaspars Ozols

Appl. Sci. 2025, 15(2), 647; https://doi.org/10.3390/app15020647 - 10 Jan 2025

Abstract

Differential privacy has recently gained prominence, especially in the context of private machine learning. While the definition of differential privacy makes it possible to provably limit the amount of information leaked by an algorithm, practical implementations of differentially private algorithms often contain subtle [...] Read more.

Differential privacy has recently gained prominence, especially in the context of private machine learning. While the definition of differential privacy makes it possible to provably limit the amount of information leaked by an algorithm, practical implementations of differentially private algorithms often contain subtle vulnerabilities. Therefore, there is a need for effective methods that can audit differentially private algorithms before they are deployed in the real world. The article examines studies that recommend privacy guarantees for differential private machine learning. It covers a wide range of topics on the subject and provides comprehensive guidance for privacy auditing schemes based on privacy attacks to protect machine-learning models from privacy leakage. Our results contribute to the growing literature on differential privacy in the realm of privacy auditing and beyond and pave the way for future research in the field of privacy-preserving models. Full article

22 pages, 11309 KiB

Open AccessArticle

Exploring Adsorption Performance of Functionalized Mesoporous Silicas with a Different Pore Structure as Strong Cation-Exchange Sorbents for Solid-Phase Extraction of Atropine and Scopolamine

by Fernando L. Vera-Baquero, Sonia Morante-Zarcero, Damián Pérez-Quintanilla and Isabel Sierra

Appl. Sci. 2025, 15(2), 646; https://doi.org/10.3390/app15020646 - 10 Jan 2025

Abstract

In this work, mesoporous silicas with two types of mesoporous structures were synthesized and functionalized with sulfonic acid groups: MCM-41-SO₃H (honeycomb-like hexagonal structure) and MSU-2-SO₃H (three-dimensional porous structure with wormhole pores). The synthesized materials were characterized by X-ray diffraction, [...] Read more.

In this work, mesoporous silicas with two types of mesoporous structures were synthesized and functionalized with sulfonic acid groups: MCM-41-SO₃H (honeycomb-like hexagonal structure) and MSU-2-SO₃H (three-dimensional porous structure with wormhole pores). The synthesized materials were characterized by X-ray diffraction, transmission electron microscopy, scanning electron microscopy, nitrogen adsorption–desorption, Fourier-transform infrared spectroscopy, ²⁹Si solid-state nuclear magnetic resonance spectroscopy, and elemental analysis. The obtained functionalized materials were evaluated as sorbents for strong cation-exchange solid-phase extraction (SPE) to determine their efficiency in the adsorption and desorption of tropane alkaloids (atropine and scopolamine). The loading solvents, loading volume, analyte concentration, and elution volume were studied, using 50 mg of both materials. Analyses were carried out by ultra-high performance liquid chromatography coupled to triple quadrupole tandem mass spectrometry. The synthesized MCM-41-SO₃H material presented the highest recovery efficiency and has proven to be a promising sorbent for strong cation-exchange SPE of atropine and scopolamine in aqueous media. The high degree of functionalization of MCM-41-SO₃H and the high accessibility of the sulfonic groups for the target analytes, due to the regularity and uniformity of their pores, maximize the contact between the alkaloids and the sorbent, favoring efficient adsorption. Full article

(This article belongs to the Special Issue New Trends in Functional and Multifunctional Advanced Materials)

► Show Figures

Figure 1

31 pages, 1054 KiB

Open AccessReview

Photoconductive Semiconductor Switches: Materials, Physics, and Applications

by Vincent Meyers, Lars Voss, Jack D. Flicker, Luciano Garcia Rodriguez, Harold P. Hjalmarson, Jane Lehr, Nicolas Gonzalez, Gregory Pickrell, Soroush Ghandiparsi and Robert Kaplar

Appl. Sci. 2025, 15(2), 645; https://doi.org/10.3390/app15020645 - 10 Jan 2025

Abstract

Photoconductive semiconductor switching (PCSS) devices have unique characteristics to address the growing need for electrically isolated, optically gated, picosecond-scale jitter devices capable of operating at high voltage, current, and frequency. The state of the art in material selection, doping, triggering, and system integration [...] Read more.

Photoconductive semiconductor switching (PCSS) devices have unique characteristics to address the growing need for electrically isolated, optically gated, picosecond-scale jitter devices capable of operating at high voltage, current, and frequency. The state of the art in material selection, doping, triggering, and system integration in PCSSs is presented. The material properties and doping considerations of GaN, GaAs, SiC, diamond, and β-Ga₂O₃ in the fabrication of PCSS devices are discussed. A review of the current understanding of the physics of the high-gain mode known as lock-on is presented. Full article

(This article belongs to the Special Issue High-Voltage Technology and Engineering Enabling Life and Physical Science Advancement)

21 pages, 824 KiB

Open AccessArticle

Effect of Precision Nitrogen Fertilization of Grassland on Soil Microbial Structure

by Anna Miernik and Jerzy Korończok

Appl. Sci. 2025, 15(2), 644; https://doi.org/10.3390/app15020644 - 10 Jan 2025

Abstract

The synergistic application of advanced technologies enables precise determination of plant growth, health, and nutritional requirements. However, despite the widespread use of modern technologies, the microbial status of the soil is often neglected, even though it significantly impacts soil productivity. Soil microbial activity [...] Read more.

The synergistic application of advanced technologies enables precise determination of plant growth, health, and nutritional requirements. However, despite the widespread use of modern technologies, the microbial status of the soil is often neglected, even though it significantly impacts soil productivity. Soil microbial activity serves as a crucial indicator of site-specific soil conditions. This article presents efforts to explore the quantitative and qualitative relationships between identified actinomycetes and soil nitrate content, as well as their distribution within the soil profile. Field data analysis facilitated the assessment of nitrate concentrations and the evaluation of the quantitative and species composition of actinomycetes in the soil profile at depths ranging from 0.05 to 0.35 m. The highest nitrate concentration (22 mg/100 g of soil) and actinomycete abundance (1076 CFU/g of soil) were observed in the topsoil layer. Additionally, spatial correlations between these parameters were analyzed for each soil layer. The correlation coefficients were approximately −0.6, indicating an inverse relationship. Areas with low nitrogen content corresponded to reduced microbial abundance within the soil profile, as supported by the spatial correlation data. These findings demonstrate the potential to predict actinomycete abundance in the soil profile based on nitrate content, offering valuable insights into soil health and productivity. Full article

(This article belongs to the Special Issue Engineering of Smart Agriculture—2nd Edition)

22 pages, 1648 KiB

Open AccessArticle

Hake Fish Preservation Using Plant-Based Impregnated Polylactic Acid Food Films as Active Packaging

by Fini Sánchez-García, Noelia D. Machado, María Tirado-Fernández, Cristina Cejudo-Bastante, Ana M. Roldán, Casimiro Mantell-Serrano and Lourdes Casas-Cardoso

Appl. Sci. 2025, 15(2), 643; https://doi.org/10.3390/app15020643 - 10 Jan 2025

Abstract

Global fish consumption has steadily increased; however, fishery products are difficult to preserve. Active packaging has emerged as an alternative to improve its conservation. In this work, fresh hake fillets were packaged in commercial polylactic acid films impregnated with olive leaf extract using [...] Read more.

Global fish consumption has steadily increased; however, fishery products are difficult to preserve. Active packaging has emerged as an alternative to improve its conservation. In this work, fresh hake fillets were packaged in commercial polylactic acid films impregnated with olive leaf extract using supercritical CO₂. The impregnation was performed at 35 °C and 400 bar for 1 h. The ABTS assay was used to determine the antioxidant activity, and migration tests were performed using food simulants A and D2 for 10 days at 5 °C. The fresh fillets were packaged in impregnated and control films and stored for 12 days at 4 °C. The microbiological, physical (drip loss, a_w, pH, and color) and chemical parameters (total volatile base and trimethylamine) were analyzed. The impregnated films presented a 706 μg extract mg⁻¹ polymer, showing a 2-fold extract release using food simulant D2 than simulant A. After hake storage using impregnated films, reduced microbial count, and drip loss, maintaining the pH stability was obtained. The color turned yellowish and no detectable olfactory presence of the extract was noted. The chemical parameters were similar in both types of films. The proposed biodegradable packaging with olive by-products preserves moisture and controls microbial growth, representing an eco-friendly alternative. Full article

(This article belongs to the Special Issue Applications of Analytical Chemistry in Food Science)

► Show Figures

Figure 1

24 pages, 3627 KiB

Open AccessArticle

Strategic Resource Planning for Sustainable Biogas Integration in Hybrid Renewable Energy Systems

by Pooriya Motevakel, Carlos Roldán-Blay, Carlos Roldán-Porta, Guillermo Escrivá-Escrivá and Daniel Dasí-Crespo

Appl. Sci. 2025, 15(2), 642; https://doi.org/10.3390/app15020642 - 10 Jan 2025

Abstract

In response to the growing demand for sustainable energy and the environmental impacts of fossil fuels, renewable sources like biomass have become crucial, especially in regions rich in agricultural and animal waste. This study focuses on a real-life project in Aras de los [...] Read more.

In response to the growing demand for sustainable energy and the environmental impacts of fossil fuels, renewable sources like biomass have become crucial, especially in regions rich in agricultural and animal waste. This study focuses on a real-life project in Aras de los Olmos, Spain, where solar, wind, and biogas from biomass serve as primary energy sources, supplemented by a hydro-based storage system to stabilize supply. Central to the research is optimizing biomass inflow to the biogas reactor—the primary controllable variable—to effectively manage the supply chain, maximize energy output, and minimize logistical costs. The study addresses practical challenges by utilizing real data on demand, truck capacities, and costs and employing robust optimization tools like Gurobi. It demonstrates how optimized biomass flow can secure energy needs during high demand or when other renewables are unavailable. Integrating technical and economic aspects, it offers a comprehensive and practical model for sustainable and economically viable energy production in rural communities. It provides a foundational framework for future renewable energy and optimized energy storage system studies. Full article

(This article belongs to the Special Issue Advances in the Sustainability and Energy Efficiency of Buildings)

► Show Figures

Figure 1

17 pages, 3767 KiB

Open AccessArticle

PGN: Progressively Guided Network with Pixel-Wise Attention for Underwater Image Enhancement

by Huidi Jia, Qiang Wang, Bo Fu, Zhimin Zheng and Yandong Tang

Appl. Sci. 2025, 15(2), 641; https://doi.org/10.3390/app15020641 - 10 Jan 2025

Abstract

Light scattering and attenuation in water degrade underwater images with low visibility and color distortion, which often interfere with the high-level visual tasks of underwater autonomous robots. Most existing deep learning methods for underwater image enhancement only supervise the final output of network [...] Read more.

Light scattering and attenuation in water degrade underwater images with low visibility and color distortion, which often interfere with the high-level visual tasks of underwater autonomous robots. Most existing deep learning methods for underwater image enhancement only supervise the final output of network and ignore the promotion effect of the intermediate results on the final feature representation. These supervision methods affect the feature representation ability, network efficiency, and ability. In this paper, we present a novel idea of multiple-stage supervision to guide the network to learn useful features correctly and progressively. With this idea, we propose a pixel-wise Progressive Guided Network (PGN) for underwater image enhancement to take advantage of the network’s intermediate results and promote the final enhancement effect. The Pixel-Wise Attention Module is designed by introducing supervision in each stage to progressively promote the representation ability of the features and the recovered image quality. The experimental results on several datasets demonstrate that our method outperforms recent state-of-the-art underwater image enhancement methods. Full article

► Show Figures

Figure 1

23 pages, 9093 KiB

Open AccessArticle

Generating Input Ground Motions for Seismic Risk Assessment Using Recorded Ground Motions from the Moderate Magnitude Earthquake

by Hae Yeon Ji and Jung Han Kim

Appl. Sci. 2025, 15(2), 640; https://doi.org/10.3390/app15020640 - 10 Jan 2025

Abstract

To secure the seismic performance of structures, seismic risk assessment is necessary to quantify safety against beyond-design-based earthquakes and seismic design. For the seismic risk assessment of structures, the input ground motions corresponding to the seismic intensity for evaluation are required as seismic [...] Read more.

To secure the seismic performance of structures, seismic risk assessment is necessary to quantify safety against beyond-design-based earthquakes and seismic design. For the seismic risk assessment of structures, the input ground motions corresponding to the seismic intensity for evaluation are required as seismic loads, which must reflect the tectonic characteristics and site conditions. In this study, ground motions recorded in regions of low to moderate seismicity were used to generate examples of input ground motions for seismic risk assessment. A uniform hazard spectrum (UHS) was used as the target spectrum for risk assessment, following the guidelines. The magnitude and distance parameters of the scenario earthquake for seismic risk assessment were determined via hazard de-aggregation. The empirical Green’s function method (EGFM) was used to match the ground motion recorded at the site with the seismic intensity required for seismic risk assessment. In addition, a spectral matching process was applied to ensure that the input ground motion was compatible with the response spectrum used in seismic risk assessment. In this process, the convergence characteristics of the spectral matching to the target spectrum were analyzed. Consequently, the spectral conditions for selecting the ground motion for the seismic risk assessment were determined. Full article

18 pages, 621 KiB

Open AccessArticle

Use of Hydroxyapatite Nanoparticles to Reduce Cd Contamination in Agricultural Soils: Effects on Growth and Development of Chenopodium quinoa Willd

by Rocío González-Feijoo, Cecilia Martínez-Castillo, Vanesa Santás-Miguel, Daniel Arenas-Lago and Paula Pérez-Rodríguez

Appl. Sci. 2025, 15(2), 639; https://doi.org/10.3390/app15020639 - 10 Jan 2025

Abstract

Soil contamination and degradation have prompted extensive research into remediation techniques. A promising approach involves the use of nanoparticles, which can mitigate heavy metal contamination, such as cadmium (Cd), without adversely affecting crop development. This study evaluated the effects of hydroxyapatite nanoparticles (HANPs) [...] Read more.

Soil contamination and degradation have prompted extensive research into remediation techniques. A promising approach involves the use of nanoparticles, which can mitigate heavy metal contamination, such as cadmium (Cd), without adversely affecting crop development. This study evaluated the effects of hydroxyapatite nanoparticles (HANPs) on the growth of Chenopodium quinoa Willd. in soils contaminated with varying Cd contents (0, 5, 10, 25, and 50 mg/kg). The results indicated that HANPs enhanced early shoot development, particularly in soils with Cd concentrations ≥10 mg/kg, while significantly reducing Cd accumulation in plant tissues. HANPs also decreased Cd mobility in soil, retaining it in fractions less available for plant uptake. Germination rates differed between pot experiments and phytotoxicity assays, although the first demonstrated greater Cd immobilization in HANP-treated soils, resulting in reduced Cd translocation to quinoa shoots. These findings highlight the potential of HANPs as an effective tool for remediating Cd-contaminated soils, thereby improving soil pollution, crop safety, and quality for human consumption. Full article

(This article belongs to the Special Issue Pollution Control and Environmental Remediation)

► Show Figures

Figure 1

18 pages, 1761 KiB

Open AccessArticle

Computer Vision-Based Drowsiness Detection Using Handcrafted Feature Extraction for Edge Computing Devices

by Valerius Owen and Nico Surantha

Appl. Sci. 2025, 15(2), 638; https://doi.org/10.3390/app15020638 - 10 Jan 2025

Abstract

Drowsy driving contributes to over 6000 fatal incidents annually in the US, underscoring the need for effective, non-intrusive drowsiness detection. This study seeks to address detection challenges, particularly in non-standard head positions. Our innovative approach leverages computer vision by combining facial feature detection [...] Read more.

Drowsy driving contributes to over 6000 fatal incidents annually in the US, underscoring the need for effective, non-intrusive drowsiness detection. This study seeks to address detection challenges, particularly in non-standard head positions. Our innovative approach leverages computer vision by combining facial feature detection using Dlib, head pose estimation with the HOPEnet model, and analyses of the percentage of eyelid closure over time (PERCLOS) and the percentage of mouth opening over time (POM). These are integrated with traditional machine learning models, such as Support Vector Machines, Random Forests, and XGBoost. These models were chosen for their ability to process detailed information from facial landmarks, head poses, PERCLOS, and POM. They achieved a high overall accuracy of 86.848% in detecting drowsiness, with a small overall model size of 5.05 MB and increased computational efficiency. The models were trained on the National Tsing Hua University Driver Drowsiness Detection Dataset, making them highly suitable for devices with a limited computational capacity. Compared to the baseline model from the literature, which achieved an accuracy of 84.82% and a larger overall model size of 37.82 MB, the method proposed in this research shows a notable improvement in the efficiency of the model with relatively similar accuracy. These findings provide a framework for future studies, potentially improving sleepiness detection systems and ultimately saving lives by enhancing road safety. Full article

(This article belongs to the Section Computing and Artificial Intelligence)

► Show Figures

Figure 1

12 pages, 440 KiB

Open AccessArticle

The Limited Impact of Running Performance on Football Success in the Turkish Super League

by Spyridon Plakias, Sotiris Tasoulis, Angelos E. Kyranoudis, Christos Kokkotis and Serafeim Moustakidis

Appl. Sci. 2025, 15(2), 637; https://doi.org/10.3390/app15020637 - 10 Jan 2025

Abstract

Given that performance in football depends on tactical, technical, physical, and mental skills, the purpose of this study was to investigate whether there are differences in running performance between winning and non-winning teams in the Turkish League, taking into account the influence of [...] Read more.

Given that performance in football depends on tactical, technical, physical, and mental skills, the purpose of this study was to investigate whether there are differences in running performance between winning and non-winning teams in the Turkish League, taking into account the influence of game location and the comparative quality of the team and its opponents. Utilizing a dataset from the 2021–2022 season provided by InStat Fitness, an optical tracking technology platform certified by FIFA, the analysis included 185 matches after adjusting for matches with red card incidents. The research employed both two-way ANCOVA and binary logistic regression analyses to explore the relationships between running performance (categorized into four intensity zones) and match results, considering factors such as match location and teams’ strength. The results of the two-way ANCOVAs indicate that running performance metrics, specifically the distances covered at different intensities, even in cases where statistically significant differences are observed, have small practical significance (partial eta squared ≤ 0.03 in all cases). Conversely, as shown by the binary logistic regression, home advantage triples the probability of winning (p < 0.001, Exp(B) = 3.119), while the increase in probability caused by team quality (p < 0.001, Exp(B) = 1.085) and the decrease caused by opponent quality (p < 0.001, Exp(B) = 0.911) are also significant. The conclusions highlight that running performance metrics are not decisive predictors of match outcomes in professional football. This suggests the importance of integrating tactical, technical, and psychological factors into team preparation and performance analysis. This study underscores the need for future research to adopt dynamic methods that reflect the game’s fluid nature and to explore these relationships across various leagues and seasons to enhance the generalizability of the findings. Full article

(This article belongs to the Special Issue Human Performance in Sports and Training)

19 pages, 1976 KiB

Open AccessArticle

A Social Media Dataset and H-GNN-Based Contrastive Learning Scheme for Multimodal Sentiment Analysis

by Jiao Peng, Yue He, Yongjuan Chang, Yanyan Lu, Pengfei Zhang, Zhonghong Ou and Qingzhi Yu

Appl. Sci. 2025, 15(2), 636; https://doi.org/10.3390/app15020636 - 10 Jan 2025

Abstract

Multimodal sentiment analysis faces a number of challenges, including modality missing, modality heterogeneity gap, incomplete datasets, etc. Previous studies usually adopt schemes like meta-learning or multi-layer structures. Nevertheless, these methods lack interpretability for the interaction between modalities. In this paper, we constructed a [...] Read more.

Multimodal sentiment analysis faces a number of challenges, including modality missing, modality heterogeneity gap, incomplete datasets, etc. Previous studies usually adopt schemes like meta-learning or multi-layer structures. Nevertheless, these methods lack interpretability for the interaction between modalities. In this paper, we constructed a new dataset, SM-MSD, for sentiment analysis in social media (SAS) that differs significantly from conventional corpora, comprising 10K instances of diverse data from Twitter, encompassing text, emoticons, emojis, and text embedded in images. This dataset aims to reflect authentic social scenarios and various emotional expressions, and provides a meaningful and challenging evaluation benchmark for multimodal sentiment analysis in specific contexts. Furthermore, we propose a multi-task framework based on heterogeneous graph neural networks (H-GNNs) and contrastive learning. For the first time, heterogeneous graph neural networks are applied to multimodal sentiment analysis tasks. In the case of additional labeling data, it guides the emotion prediction of the missing mode. We conduct extensive experiments on multiple datasets to verify the effectiveness of the proposed scheme. Experimental results demonstrate that our proposed scheme surpasses state-of-the-art methods by 1.7% and 0 in accuracy and 1.54% and 4.9% in F1-score on the MOSI and MOSEI datasets, respectively, and exhibits robustness to modality missing scenarios. Full article

► Show Figures

Figure 1

28 pages, 7902 KiB

Open AccessArticle

A Case Study of Using Numerical Analysis to Assess the Slope Stability of National Freeways in Northern Taiwan

by Hao-Wei Chiu, Yi-Hao Tsai, Chao-Wei Tang, Chih-Yu Chu and Shong-Loong Chen

Appl. Sci. 2025, 15(2), 635; https://doi.org/10.3390/app15020635 - 10 Jan 2025

Abstract

Taiwan is located at a junction of tectonic plates, which results in frequent earthquakes. Its terrain is mostly hilly, and its rainfall ranks among the highest in the world. Each of these elements affects the stability of slopes in various regions of Taiwan. [...] Read more.

Taiwan is located at a junction of tectonic plates, which results in frequent earthquakes. Its terrain is mostly hilly, and its rainfall ranks among the highest in the world. Each of these elements affects the stability of slopes in various regions of Taiwan. Several slopes along Taiwan’s Freeway 1 and 5 have experienced landslides and rockfalls. It is imperative that the slope stability of these national freeways be analyzed to avoid future slope collapses brought on by precipitation or other outside factors. Thus, three sites on Taiwan’s Freeway 1 and 5 were chosen for numerical slope stability analysis in this study. PLAXIS 2D CE (Version: 24.02.00.1144) finite element software was used in this study to simulate and analyze the safety of freeway slope protection projects. Displacements induced by normal and high groundwater levels were discussed. Moreover, a pseudo-static study of slope displacements under seismic conditions was performed. According to the results of the numerical study, the force operating on the slope was centered on the sliding surface when the groundwater level was normal, and it extended to the top when the groundwater level was high. By comparison, under seismic conditions, the force acting on the slope extended to the whole slope. Furthermore, the slope safety factor of Site 1 was greater than the design specification value in three different scenarios. This confirms that the slope protection project at Site 1 is effective. Full article

18 pages, 342 KiB

Open AccessArticle

The Addition of Tomato and Spinach Powder to Semolina Pasta: A Study of the Impact of the Production Process and Cooking on Phenolic Compounds

by Silvia Marzocchi, Federica Pasini, Renzo Santi and Maria Fiorenza Caboni

Appl. Sci. 2025, 15(2), 634; https://doi.org/10.3390/app15020634 - 10 Jan 2025

Abstract

Pasta is a staple food with daily recommended consumption; thus, it can be an excellent vehicle for delivering bioactive compounds like phenols. However, the high-temperature drying process in pasta production, combined with cooking in boiling water, can significantly reduce the concentration of phenolic [...] Read more.

Pasta is a staple food with daily recommended consumption; thus, it can be an excellent vehicle for delivering bioactive compounds like phenols. However, the high-temperature drying process in pasta production, combined with cooking in boiling water, can significantly reduce the concentration of phenolic compounds. This study aimed to enhance the phenolic content of traditional semolina pasta by incorporating tomato and spinach powder into the recipe. High-performance liquid chromatography–electrospray ionization triple quadrupole mass spectrometry (HPLC-ESI-QqQ-MS) was employed to analyse the free and bound phenolic content in the raw materials, as well as in both dried uncooked and cooked pasta. The addition of tomato and spinach powders, known for their high content in bioactive compounds, increased the overall phenolic content of the final enriched pasta by three and two times, respectively, compared to the semolina and whole-wheat semolina pasta. These findings suggest that pasta enriched with tomato and spinach could serve as a functional food with a greater nutritional profile and health benefits through the enhanced delivery of phenolic compounds. Full article

21 pages, 1635 KiB

Open AccessArticle

Enhanced Bug Priority Prediction via Priority-Sensitive Long Short-Term Memory–Attention Mechanism

by Geunseok Yang, Jinfeng Ji and Jaehee Kim

Appl. Sci. 2025, 15(2), 633; https://doi.org/10.3390/app15020633 - 10 Jan 2025

Abstract

The rapid expansion of software applications has led to an increase in the frequency of bugs, which are typically reported through user-submitted bug reports. Developers prioritize these reports based on severity and project schedules. However, the manual process of assigning bug priorities is [...] Read more.

The rapid expansion of software applications has led to an increase in the frequency of bugs, which are typically reported through user-submitted bug reports. Developers prioritize these reports based on severity and project schedules. However, the manual process of assigning bug priorities is time-consuming and prone to inconsistencies. To address these limitations, this study presents a Priority-Sensitive LSTM–Attention mechanism for automating bug priority prediction. The proposed approach extracts features such as product and component details from bug repositories and preprocesses the data to ensure consistency. Priority-based feature selection is applied to align the input data with the task of bug prioritization. These features are processed through a Long Short-Term Memory (LSTM) network to capture sequential dependencies, and the outputs are further refined using an Attention mechanism to focus on the most relevant information for prediction. The effectiveness of the proposed model was evaluated using datasets from the Eclipse and Mozilla open-source projects. Compared to baseline models such as Naïve Bayes, Random Forest, Decision Tree, SVM, CNN, LSTM, and CNN-LSTM, the proposed model achieved a superior performance. It recorded an accuracy of 93.00% for Eclipse and 84.11% for Mozilla, representing improvements of 31.11% and 40.39%, respectively, over the baseline models. Statistical verification confirmed that these performance gains were significant. This study distinguishes itself by integrating priority-based feature selection with a hybrid LSTM–Attention architecture, which enhances prediction accuracy and robustness compared to existing methods. The results demonstrate the potential of this approach to streamline bug prioritization, improve project management efficiency, and assist developers in resolving high-priority issues. Full article

(This article belongs to the Section Computing and Artificial Intelligence)

25 pages, 2088 KiB

Open AccessArticle

Real-Time Optimization of RISC-V Processors Based on Branch Prediction and Division Data Dependency

by Zhiwei Jin, Hesong Di, Tingpeng Hu and Peng Wang

Appl. Sci. 2025, 15(2), 632; https://doi.org/10.3390/app15020632 - 10 Jan 2025

Abstract

Whether a processor can meet the real-time requirements of a system is a crucial factor affecting the security of real-time systems. Currently, the methods for evaluating hardware real-time performance and the quality of real-time performance are not comprehensive. To address these issues, a [...] Read more.

Whether a processor can meet the real-time requirements of a system is a crucial factor affecting the security of real-time systems. Currently, the methods for evaluating hardware real-time performance and the quality of real-time performance are not comprehensive. To address these issues, a “Hardware Real-Time Parameter (H_rtp)” is proposed, which integrates the concepts of “whether it meets real-time system requirements”, “execution speed”, and “operational stability”. This parameter is used to evaluate the real-time performance of RISC-V processors before and after optimization. The processor is optimized for real-time performance using a “simplified local history branch predictor” and a “division data dependency module”. Experimental results show that the processor’s branch prediction accuracy and division calculation speed have both improved. When running the CoreMark benchmark program and the division test program, the test results improved, indicating an enhanced “real-time” performance of the hardware. The changes in the “Standalone Hardware Real-Time Performance Parameter (H_rtp)” data are consistent with theoretical analysis, and it can meet the evaluation needs for “hardware real-time performance”. Full article

Journal Description

Applied Sciences

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.