Pub Date : 2026-05-25Epub Date: 2026-01-29DOI: 10.1016/j.eswa.2026.131296
Hansen Zhang , Miao Yang , Can Pan , Leyuan Wang , Jiaju Tao
Underwater images commonly suffer from blurring, low contrast, and color distortion due to light scattering effects and wavelength-dependent absorption. From a frequency-domain perspective, these degradations significantly impair the spectral representation: they not only attenuate high-frequency amplitudes (causing detail loss) but also weaken specific color channel energies (inducing color deviation) while reducing overall spectral energy (leading to contrast deterioration). To address these challenges, this paper proposes an innovative two-stage underwater image enhancement framework, termed APD-LDM. In the first stage, we design an Amplitude-Phase Decomposition Network (APDNet) that performs end-to-end learning on paired underwater image data to preliminarily recover amplitude information degraded by absorption and scattering effects. The second stage employs a conditional diffusion model for refined reconstruction, where latent representations of degraded images serve as conditional constraints to guide the diffusion process toward more realistic underwater image features. Additionally, we introduce a self-constrained consistency loss function to further optimize network training. Extensive experiments demonstrate that the proposed method achieves superior effectiveness and robustness in both subjective visual quality and objective metrics. The code is available at https://github.com/JOU-UIP/APD-LDM.
{"title":"Amplitude -phase decomposition-based latent diffusion model for underwater image enhancement","authors":"Hansen Zhang , Miao Yang , Can Pan , Leyuan Wang , Jiaju Tao","doi":"10.1016/j.eswa.2026.131296","DOIUrl":"10.1016/j.eswa.2026.131296","url":null,"abstract":"<div><div>Underwater images commonly suffer from blurring, low contrast, and color distortion due to light scattering effects and wavelength-dependent absorption. From a frequency-domain perspective, these degradations significantly impair the spectral representation: they not only attenuate high-frequency amplitudes (causing detail loss) but also weaken specific color channel energies (inducing color deviation) while reducing overall spectral energy (leading to contrast deterioration). To address these challenges, this paper proposes an innovative two-stage underwater image enhancement framework, termed APD-LDM. In the first stage, we design an Amplitude-Phase Decomposition Network (APDNet) that performs end-to-end learning on paired underwater image data to preliminarily recover amplitude information degraded by absorption and scattering effects. The second stage employs a conditional diffusion model for refined reconstruction, where latent representations of degraded images serve as conditional constraints to guide the diffusion process toward more realistic underwater image features. Additionally, we introduce a self-constrained consistency loss function to further optimize network training. Extensive experiments demonstrate that the proposed method achieves superior effectiveness and robustness in both subjective visual quality and objective metrics. The code is available at <span><span>https://github.com/JOU-UIP/APD-LDM</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131296"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In long-term time series forecasting (LSTF), a fundamental challenge lies in simultaneously capturing fine-grained local dynamics and long-range global dependencies within inherently complex and non-stationary temporal series. However, most existing forecasting architectures rely on single-structure paradigms, each exhibiting inherent representational biases–for example, CNNs are constrained by limited receptive fields, while Transformers often overlook fine-grained local patterns. More critically, these architectures typically operate in isolation, lacking collaborative mechanisms to effectively integrate their complementary modeling capabilities. To address these limitations, we propose MatNet, a Multi-scale Adaptive Forecasting Network with a novel bidirectional collaborative architecture designed to establish bidirectional collaborative pathways between CNN and Transformer branches. Within this architecture, local representations extracted by CNNs are leveraged to refine and enrich the global context modeled by Transformers, thereby improving the model’s sensitivity to fine-grained temporal structures. Conversely, global dependencies captured by Transformer provide high-level semantic guidance to CNNs, enabling them to focus on contextually salient local regions and enhance representation coherence. Additionally, we introduce a Dynamic Temporal-Aware Router that adaptively extracts and fuses temporal features across multiple scales, enabling adaptive multi-scale modeling. Extensive experiments on nine public datasets demonstrate that MatNet consistently outperforms existing state-of-the-art methods in forecasting accuracy.
{"title":"MatNet : Multi-scale adaptive time series forecasting network with bidirectional collaborative pathways","authors":"Guangming Zi, Yujun Zhu, Xin He, Yong Xu, Qun Fang","doi":"10.1016/j.eswa.2026.131493","DOIUrl":"10.1016/j.eswa.2026.131493","url":null,"abstract":"<div><div>In long-term time series forecasting (LSTF), a fundamental challenge lies in simultaneously capturing fine-grained local dynamics and long-range global dependencies within inherently complex and non-stationary temporal series. However, most existing forecasting architectures rely on single-structure paradigms, each exhibiting inherent representational biases–for example, CNNs are constrained by limited receptive fields, while Transformers often overlook fine-grained local patterns. More critically, these architectures typically operate in isolation, lacking collaborative mechanisms to effectively integrate their complementary modeling capabilities. To address these limitations, we propose MatNet, a Multi-scale Adaptive Forecasting Network with a novel bidirectional collaborative architecture designed to establish bidirectional collaborative pathways between CNN and Transformer branches. Within this architecture, local representations extracted by CNNs are leveraged to refine and enrich the global context modeled by Transformers, thereby improving the model’s sensitivity to fine-grained temporal structures. Conversely, global dependencies captured by Transformer provide high-level semantic guidance to CNNs, enabling them to focus on contextually salient local regions and enhance representation coherence. Additionally, we introduce a Dynamic Temporal-Aware Router that adaptively extracts and fuses temporal features across multiple scales, enabling adaptive multi-scale modeling. Extensive experiments on nine public datasets demonstrate that MatNet consistently outperforms existing state-of-the-art methods in forecasting accuracy.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131493"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-05DOI: 10.1016/j.eswa.2026.131532
Hongmei Li , Dongxia Qu , Taibo Luo
Natural disasters occur frequently in the real world, and the secondary disasters they trigger can also result in significant losses. Well-organized disaster response is essential for providing timely, effective medical services in such emergencies. This paper addresses a post-disaster emergency response problem involving the determination of the location and number of temporary medical centers (TMCs) and the planning of casualty transportation, considering both primary and secondary disasters under uncertainty. A two-stage robust optimization model based on the minimax regret criterion is proposed. In the first stage, the locations and allocations for primary disaster casualties are determined prior to the occurrence of secondary disasters. In the second stage, casualties resulting from secondary disasters are transported to the established TMCs. To handle uncertainty in the number and spatial distribution of secondary disaster casualties, a minimax regret approach is employed across a set of predefined scenarios. To enhance computational efficiency, a scenario relaxation algorithm based on row generation is developed. Case studies based on the Lushan Earthquake are conducted to validate the feasibility and effectiveness of the model. Results demonstrate that incorporating secondary disasters significantly improves the efficiency of casualty treatment compared to models considering only primary disasters. Under the uncertainty of secondary disasters, constructing a limited number of critical scenarios is sufficient, and large-capacity TMCs are more recommended.
{"title":"A robust model on the location of temporary medical centers considering secondary disasters","authors":"Hongmei Li , Dongxia Qu , Taibo Luo","doi":"10.1016/j.eswa.2026.131532","DOIUrl":"10.1016/j.eswa.2026.131532","url":null,"abstract":"<div><div>Natural disasters occur frequently in the real world, and the secondary disasters they trigger can also result in significant losses. Well-organized disaster response is essential for providing timely, effective medical services in such emergencies. This paper addresses a post-disaster emergency response problem involving the determination of the location and number of temporary medical centers (TMCs) and the planning of casualty transportation, considering both primary and secondary disasters under uncertainty. A two-stage robust optimization model based on the minimax regret criterion is proposed. In the first stage, the locations and allocations for primary disaster casualties are determined prior to the occurrence of secondary disasters. In the second stage, casualties resulting from secondary disasters are transported to the established TMCs. To handle uncertainty in the number and spatial distribution of secondary disaster casualties, a minimax regret approach is employed across a set of predefined scenarios. To enhance computational efficiency, a scenario relaxation algorithm based on row generation is developed. Case studies based on the Lushan Earthquake are conducted to validate the feasibility and effectiveness of the model. Results demonstrate that incorporating secondary disasters significantly improves the efficiency of casualty treatment compared to models considering only primary disasters. Under the uncertainty of secondary disasters, constructing a limited number of critical scenarios is sufficient, and large-capacity TMCs are more recommended.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131532"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192699","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-01-31DOI: 10.1016/j.eswa.2026.131436
Changhong Tai , Huiying Jin , Qi Wang , Hai Dong , Pengcheng Zhang
This paper addresses the challenges of workflow scheduling in hybrid cloud environments, where unknown task execution times and resource demand deviations often lead to prolonged makespan and scheduling failures. We propose FBiLSTM-DPBWS, a novel budget-constrained workflow scheduling algorithm. The core contributions of this algorithm are reflected at two levels: Firstly, a novel FBiLSTM regression prediction model is proposed. By integrating the Flash Attention mechanism with bidirectional long short-term memory networks, it can accurately and synchronously predict the instruction counts of all subtasks based solely on the directed acyclic graph structure and prior information about the workflow before task execution, thereby estimating their execution times on heterogeneous resources. This fundamentally addresses the challenge of unknown execution times. Secondly, a dynamic critical-path-priority-based budget-constrained workflow scheduling algorithm, DPBWS, is designed. Instead of simply combining virtual machine and serverless function resources, this algorithm dynamically adjusts the priority of critical-path tasks. It adaptively selects the optimal resource type based on predicted instruction counts, real-time budget consumption, and the task’s compute or data-intensive characteristics. The algorithm explicitly accounts for the fundamental billing differences between these resource types (whole-unit rental per billing period vs. fine-grained billing based on actual execution time), thereby minimizing makespan and maximizing scheduling success rates under limited budgets. Experiments conducted on six real-world and five large-scale synthetic datasets demonstrate that the FBiLSTM model achieves prediction accuracies ranging from 97.80% to 99.62%. Under the same budget, DPBWS significantly reduces makespan compared to the best-performing baseline and achieves scheduling success rates from 98.62% to 100% across all datasets. These results confirm the superiority, robustness, and practical applicability of the proposed method in hybrid cloud environments.
{"title":"Budget-constrained workflow scheduling using task prediction in hybrid environments","authors":"Changhong Tai , Huiying Jin , Qi Wang , Hai Dong , Pengcheng Zhang","doi":"10.1016/j.eswa.2026.131436","DOIUrl":"10.1016/j.eswa.2026.131436","url":null,"abstract":"<div><div>This paper addresses the challenges of workflow scheduling in hybrid cloud environments, where unknown task execution times and resource demand deviations often lead to prolonged makespan and scheduling failures. We propose FBiLSTM-DPBWS, a novel budget-constrained workflow scheduling algorithm. The core contributions of this algorithm are reflected at two levels: Firstly, a novel FBiLSTM regression prediction model is proposed. By integrating the Flash Attention mechanism with bidirectional long short-term memory networks, it can accurately and synchronously predict the instruction counts of all subtasks based solely on the directed acyclic graph structure and prior information about the workflow before task execution, thereby estimating their execution times on heterogeneous resources. This fundamentally addresses the challenge of unknown execution times. Secondly, a dynamic critical-path-priority-based budget-constrained workflow scheduling algorithm, DPBWS, is designed. Instead of simply combining virtual machine and serverless function resources, this algorithm dynamically adjusts the priority of critical-path tasks. It adaptively selects the optimal resource type based on predicted instruction counts, real-time budget consumption, and the task’s compute or data-intensive characteristics. The algorithm explicitly accounts for the fundamental billing differences between these resource types (whole-unit rental per billing period vs. fine-grained billing based on actual execution time), thereby minimizing makespan and maximizing scheduling success rates under limited budgets. Experiments conducted on six real-world and five large-scale synthetic datasets demonstrate that the FBiLSTM model achieves prediction accuracies ranging from 97.80% to 99.62%. Under the same budget, DPBWS significantly reduces makespan compared to the best-performing baseline and achieves scheduling success rates from 98.62% to 100% across all datasets. These results confirm the superiority, robustness, and practical applicability of the proposed method in hybrid cloud environments.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131436"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-02DOI: 10.1016/j.eswa.2026.131510
Jifei Hu , Yanli Li , Huayong Xie , Lijun Xu , Hang Zhang , Xinqiang Zhou
Federated learning (FL) enables distributed model training while preserving data privacy, but it still faces challenges from data heterogeneity and privacy constraints. Existing optimization methods aiming to flatten loss landscapes often fail to resolve inconsistencies between local and global flatness due to hyperparameter dependencies and centralized aggregation. Moreover, privacy-preserving techniques such as differential privacy (DP) can distort gradients, leading to sharper loss landscapes and hindered convergence. To tackle these issues, we propose DFedLSAM (Decentralized Federated Local Sharpness-Aware Minimization), a novel framework that eliminates the central server and uses the Sharpness-Aware Minimization (SAM) optimizer on the client side to maintain locally flattened loss landscapes. DFedLSAM adopts a dual-model architecture, where each client trains a sharing model for cross-client knowledge exchange and a private model updated via knowledge distillation (KD) from the sharing model’s soft logits, thereby reducing data heterogeneity and mitigating DP-induced sharpness. Building on this design, we introduce a perturbation-based SAM variant, integrated into the framework as DFedLSAM-Pert, which allocates perturbations according to layer-wise sensitivity and improves generalization without sacrificing privacy. Extensive experiments on benchmark image datasets and real-world medical datasets show that DFedLSAM and its perturbation-based extension DFedLSAM-Pert outperform existing baselines, especially in non-IID settings and under strict privacy budgets. These results indicate that DFedLSAM and DFedLSAM-Pert provide practical solutions for privacy-preserving FL in healthcare and other sensitive application domains.
{"title":"Local sharpness aware minimization in decentralized federated learning with privacy protection","authors":"Jifei Hu , Yanli Li , Huayong Xie , Lijun Xu , Hang Zhang , Xinqiang Zhou","doi":"10.1016/j.eswa.2026.131510","DOIUrl":"10.1016/j.eswa.2026.131510","url":null,"abstract":"<div><div>Federated learning (FL) enables distributed model training while preserving data privacy, but it still faces challenges from data heterogeneity and privacy constraints. Existing optimization methods aiming to flatten loss landscapes often fail to resolve inconsistencies between local and global flatness due to hyperparameter dependencies and centralized aggregation. Moreover, privacy-preserving techniques such as differential privacy (DP) can distort gradients, leading to sharper loss landscapes and hindered convergence. To tackle these issues, we propose DFedLSAM (Decentralized Federated Local Sharpness-Aware Minimization), a novel framework that eliminates the central server and uses the Sharpness-Aware Minimization (SAM) optimizer on the client side to maintain locally flattened loss landscapes. DFedLSAM adopts a dual-model architecture, where each client trains a sharing model for cross-client knowledge exchange and a private model updated via knowledge distillation (KD) from the sharing model’s soft logits, thereby reducing data heterogeneity and mitigating DP-induced sharpness. Building on this design, we introduce a perturbation-based SAM variant, integrated into the framework as DFedLSAM-Pert, which allocates perturbations according to layer-wise sensitivity and improves generalization without sacrificing privacy. Extensive experiments on benchmark image datasets and real-world medical datasets show that DFedLSAM and its perturbation-based extension DFedLSAM-Pert outperform existing baselines, especially in non-IID settings and under strict privacy budgets. These results indicate that DFedLSAM and DFedLSAM-Pert provide practical solutions for privacy-preserving FL in healthcare and other sensitive application domains.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131510"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122658","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Graphs have been widely used across scientific disciplines, from sociology to biology, particularly when modeling temporal evolution. Although many algorithms have been developed to discover patterns in graphs, they face three main limitations. First, most algorithms assume that each node or edge is associated with a single attribute, whereas real-world applications often involve multiple attributes to capture events more comprehensively. Second, existing methods typically require tuning several hyperparameters, which can vary significantly across different datasets. Third, most approaches focus on identifying frequent patterns, often overlooking rare but meaningful ones. To address these limitations, this paper proposes a framework for discovering anomalous sequences in attributed graphs. Instead of relying on frequency-based measures, the framework adopts an entropy-based method for pattern mining, thereby requiring at most one hyperparameter. Experimental results on real-world datasets demonstrate the effectiveness of the proposed approach in detecting anomalous sequences. Moreover, we extend the framework to applications in optics, where it is used to evaluate phase differences.
{"title":"Discovering anomalous sequences in attributed graphs: A parameter-light approach","authors":"Cheng He , Xinyang Chen , Amaël Broustet , Guoting Chen","doi":"10.1016/j.eswa.2026.131467","DOIUrl":"10.1016/j.eswa.2026.131467","url":null,"abstract":"<div><div>Graphs have been widely used across scientific disciplines, from sociology to biology, particularly when modeling temporal evolution. Although many algorithms have been developed to discover patterns in graphs, they face three main limitations. First, most algorithms assume that each node or edge is associated with a single attribute, whereas real-world applications often involve multiple attributes to capture events more comprehensively. Second, existing methods typically require tuning several hyperparameters, which can vary significantly across different datasets. Third, most approaches focus on identifying frequent patterns, often overlooking rare but meaningful ones. To address these limitations, this paper proposes a framework for discovering anomalous sequences in attributed graphs. Instead of relying on frequency-based measures, the framework adopts an entropy-based method for pattern mining, thereby requiring at most one hyperparameter. Experimental results on real-world datasets demonstrate the effectiveness of the proposed approach in detecting anomalous sequences. Moreover, we extend the framework to applications in optics, where it is used to evaluate phase differences.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131467"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122664","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-05DOI: 10.1016/j.eswa.2026.131367
Huilin Ge , Meng Li , Guanghui Wen , Yu Lu
Autonomous surface vehicles (ASVs) require reliable safety assurance to operate in complex and dynamic marine environments. This paper develops an integrated decision-support framework that couples hybrid graph-based path planning with formal verification to guarantee safe and reachable navigation. A composite roadmap is generated using the proposed HVV-E planner, which combines Voronoi-based global planning and visibility-graph refinement to produce collision-free and energy-aware trajectories. To ensure the trustworthiness of candidate routes, the navigation process is abstracted into a Linearly Priced Timed Automata (LPTA) model and formally verified using UPPAAL. Both qualitative properties, including obstacle avoidance and deadlock-freeness, and quantitative mission constraints, such as bounded travel time and energy consumption, are examined. The verification results provide explicit explanations of why a route is safe or unsafe, enabling early identification of infeasible or risky mission configurations. Experiments conducted in realistic Singapore Strait scenarios demonstrate that the proposed framework delivers transparent, safety-assured, and energy-aware navigation support for real-world ASV missions. The results highlight the value of integrating formal reasoning with intelligent planning to advance explainable and trustworthy autonomous maritime systems.
{"title":"Safety-assured decision support for ASV navigation via hybrid graph planning and timed automata verification","authors":"Huilin Ge , Meng Li , Guanghui Wen , Yu Lu","doi":"10.1016/j.eswa.2026.131367","DOIUrl":"10.1016/j.eswa.2026.131367","url":null,"abstract":"<div><div>Autonomous surface vehicles (ASVs) require reliable safety assurance to operate in complex and dynamic marine environments. This paper develops an integrated decision-support framework that couples hybrid graph-based path planning with formal verification to guarantee safe and reachable navigation. A composite roadmap is generated using the proposed HVV-E planner, which combines Voronoi-based global planning and visibility-graph refinement to produce collision-free and energy-aware trajectories. To ensure the trustworthiness of candidate routes, the navigation process is abstracted into a Linearly Priced Timed Automata (LPTA) model and formally verified using UPPAAL. Both qualitative properties, including obstacle avoidance and deadlock-freeness, and quantitative mission constraints, such as bounded travel time and energy consumption, are examined. The verification results provide explicit explanations of why a route is safe or unsafe, enabling early identification of infeasible or risky mission configurations. Experiments conducted in realistic Singapore Strait scenarios demonstrate that the proposed framework delivers transparent, safety-assured, and energy-aware navigation support for real-world ASV missions. The results highlight the value of integrating formal reasoning with intelligent planning to advance explainable and trustworthy autonomous maritime systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131367"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-05DOI: 10.1016/j.eswa.2026.131458
Meng Zhang, Jianghua Zhang, Dongchen Gao, Weibo Liu
Accurate and timely admission prediction in emergency departments is essential for improving resource allocation, enhancing patient outcomes, and mitigating overcrowding. However, the progression of emergency patients often exhibits strong temporal dynamics, and clinical decisions typically involve not only admission and non-admission but also an intermediate state of wait-and-see. To address this challenge, this study proposes a novel temporal three-way decision-making method that integrates Temporal Feature-based Multigranulation Neighborhood Rough Set (TMNRS) with Gaussian Mixture-Hidden Markov Model (GMM-HMM). Specifically, TMNRS is utilized to quantify and characterize the initial distribution of patient states from both theoretical and data-driven perspectives, thereby providing parameter support for subsequent modeling. Building on this foundation, GMM-HMM is employed to capture the dynamic evolution of patients’ conditions across three states over time. This integration facilitates interpretable state representation of the model. Finally, comprehensive experiments conducted on real-world clinical data, including comparisons with multiple benchmark models, demonstrate competitive and rob ust performance of the proposed approach in supporting temporal three-way admission decision-making for emergency patients.
{"title":"Temporal three-way decision-making for emergency admission integrating multigranulation neighborhood rough set with Gaussian mixture-hidden Markov model","authors":"Meng Zhang, Jianghua Zhang, Dongchen Gao, Weibo Liu","doi":"10.1016/j.eswa.2026.131458","DOIUrl":"10.1016/j.eswa.2026.131458","url":null,"abstract":"<div><div>Accurate and timely admission prediction in emergency departments is essential for improving resource allocation, enhancing patient outcomes, and mitigating overcrowding. However, the progression of emergency patients often exhibits strong temporal dynamics, and clinical decisions typically involve not only admission and non-admission but also an intermediate state of wait-and-see. To address this challenge, this study proposes a novel temporal three-way decision-making method that integrates Temporal Feature-based Multigranulation Neighborhood Rough Set (TMNRS) with Gaussian Mixture-Hidden Markov Model (GMM-HMM). Specifically, TMNRS is utilized to quantify and characterize the initial distribution of patient states from both theoretical and data-driven perspectives, thereby providing parameter support for subsequent modeling. Building on this foundation, GMM-HMM is employed to capture the dynamic evolution of patients’ conditions across three states over time. This integration facilitates interpretable state representation of the model. Finally, comprehensive experiments conducted on real-world clinical data, including comparisons with multiple benchmark models, demonstrate competitive and rob ust performance of the proposed approach in supporting temporal three-way admission decision-making for emergency patients.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131458"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-04DOI: 10.1016/j.eswa.2026.131525
Zhuhua Hu , Yifeng Sun , Yaochi Zhao , Wei Wu , Lingkai Kong , Keli Chen
Cooperating with maritime administrative departments to identify smuggling activities and enhance the control ability of nearshore vessels holds significant practical significance. However, existing research mostly relies on basic AIS data and simple features, making it difficult to deal with complex vessel behaviors. Especially when identifying covert and flexible smuggling activities, it is prone to misjudgment and has limited effectiveness. In real-world enforcement, distinguishing truly suspicious “Quick Smuggler” smuggling from benign high-speed transit requires modeling subtle, deep-level spatio-temporal cues that couple motion dynamics with external conditions (e.g., wind, wave, visibility) and context. Simple linear mappings and shallow temporal encoders often overfit speed bursts or local detours, causing elevated false alarms. By contrast, dilated-convolutional receptive fields in TCNs capture multi-scale temporal dependencies efficiently, while KAN layers provide adaptive nonlinear function bases to fit complex, locally varying trajectory patterns. This synergy is particularly suited to covert nighttime operations under shifting sea states, where genuine smuggling exhibits trajectory micro-structures and weather-conditioned behaviors that are hard to emulate by normal craft. To address these challenges, this study proposes a Multi-Source Information Fusion-based “Quick Smuggler” Smuggling Speedboat Trajectory Recognition method (MSIF-SSTR). First, we construct the HN_BF dataset, comprising real-world nighttime radar trajectories from the Qiongzhou Strait and corresponding meteorological data. Next, parallel TCN networks are employed to separately extract motion features, and meteorological features, enabling the model to better capture global temporal dependencies during feature extraction. Finally, the fused features are fed into an LSTM for classification, while a Kolmogorov-arnold networks (KAN) module replaces traditional fully connected layers to improve the representation of complex trajectory patterns. Experimental results demonstrate that MSIF-SSTR achieves F1-scores exceeding 94.2% on the HN_BF dataset, outperforming state-of-the-art methods with higher computational efficiency. Field applications confirm the model’s robustness.
{"title":"MSIF-SSTR: A “Quick smuggler” smuggling speedboat trajectory recognition method based on multi-source information fusion","authors":"Zhuhua Hu , Yifeng Sun , Yaochi Zhao , Wei Wu , Lingkai Kong , Keli Chen","doi":"10.1016/j.eswa.2026.131525","DOIUrl":"10.1016/j.eswa.2026.131525","url":null,"abstract":"<div><div>Cooperating with maritime administrative departments to identify smuggling activities and enhance the control ability of nearshore vessels holds significant practical significance. However, existing research mostly relies on basic AIS data and simple features, making it difficult to deal with complex vessel behaviors. Especially when identifying covert and flexible smuggling activities, it is prone to misjudgment and has limited effectiveness. In real-world enforcement, distinguishing truly suspicious “Quick Smuggler” smuggling from benign high-speed transit requires modeling subtle, deep-level spatio-temporal cues that couple motion dynamics with external conditions (e.g., wind, wave, visibility) and context. Simple linear mappings and shallow temporal encoders often overfit speed bursts or local detours, causing elevated false alarms. By contrast, dilated-convolutional receptive fields in TCNs capture multi-scale temporal dependencies efficiently, while KAN layers provide adaptive nonlinear function bases to fit complex, locally varying trajectory patterns. This synergy is particularly suited to covert nighttime operations under shifting sea states, where genuine smuggling exhibits trajectory micro-structures and weather-conditioned behaviors that are hard to emulate by normal craft. To address these challenges, this study proposes a Multi-Source Information Fusion-based “Quick Smuggler” Smuggling Speedboat Trajectory Recognition method (MSIF-SSTR). First, we construct the HN_BF dataset, comprising real-world nighttime radar trajectories from the Qiongzhou Strait and corresponding meteorological data. Next, parallel TCN networks are employed to separately extract motion features, and meteorological features, enabling the model to better capture global temporal dependencies during feature extraction. Finally, the fused features are fed into an LSTM for classification, while a Kolmogorov-arnold networks (KAN) module replaces traditional fully connected layers to improve the representation of complex trajectory patterns. Experimental results demonstrate that MSIF-SSTR achieves F1-scores exceeding 94.2% on the HN_BF dataset, outperforming state-of-the-art methods with higher computational efficiency. Field applications confirm the model’s robustness.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131525"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-25Epub Date: 2026-02-02DOI: 10.1016/j.eswa.2026.131390
Zhijing Yang , Weiwei Zhang , Mingliang Yang , Siyuan Peng , Yukai Shi , Junpeng Tan , Tianshui Chen , Liruo Zhong
This work aims to address a novel Customized Virtual Try-ON (Cu-VTON) task, enabling the superimposition of a specified garment onto a model that can be customized in terms of appearance, posture, and additional attributes. Compared with traditional VTON task, it enables users to tailor digital avatars to their individual preferences, thereby enhancing the virtual fitting experience with greater flexibility and engagement. To address this task, we introduce a Neural Clothing Tryer (NCT) framework, which exploits the advanced diffusion models equipped with semantic enhancement and controlling modules to better preserve semantic characterization and textural details of the garment and meanwhile facilitating the flexible editing of the model’s postures and appearances. Specifically, NCT introduces a semantic-enhanced module to take semantic descriptions of garments and utilizes a visual-language encoder to learn aligned features across modalities. The aligned features are served as condition input to the diffusion model to enhance the preservation of the garment’s semantics. Then, a semantic controlling module is designed to take the garment image, tailored posture image, and semantic description as input to maintain garment details while simultaneously editing model postures, expressions, and various attributes. Extensive experiments on the open available benchmark demonstrate the superior performance of the proposed NCT framework.
{"title":"Neural clothing tryer: Customized virtual try-on via semantic enhancement and controlling diffusion model","authors":"Zhijing Yang , Weiwei Zhang , Mingliang Yang , Siyuan Peng , Yukai Shi , Junpeng Tan , Tianshui Chen , Liruo Zhong","doi":"10.1016/j.eswa.2026.131390","DOIUrl":"10.1016/j.eswa.2026.131390","url":null,"abstract":"<div><div>This work aims to address a novel Customized Virtual Try-ON (Cu-VTON) task, enabling the superimposition of a specified garment onto a model that can be customized in terms of appearance, posture, and additional attributes. Compared with traditional VTON task, it enables users to tailor digital avatars to their individual preferences, thereby enhancing the virtual fitting experience with greater flexibility and engagement. To address this task, we introduce a Neural Clothing Tryer (NCT) framework, which exploits the advanced diffusion models equipped with semantic enhancement and controlling modules to better preserve semantic characterization and textural details of the garment and meanwhile facilitating the flexible editing of the model’s postures and appearances. Specifically, NCT introduces a semantic-enhanced module to take semantic descriptions of garments and utilizes a visual-language encoder to learn aligned features across modalities. The aligned features are served as condition input to the diffusion model to enhance the preservation of the garment’s semantics. Then, a semantic controlling module is designed to take the garment image, tailored posture image, and semantic description as input to maintain garment details while simultaneously editing model postures, expressions, and various attributes. Extensive experiments on the open available benchmark demonstrate the superior performance of the proposed NCT framework.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"312 ","pages":"Article 131390"},"PeriodicalIF":7.5,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146192843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号