首页 > 最新文献
paperpass

Applied Soft Computing最新文献

英文 中文
Multimodal transformers for image and audio polyphonic music transcription 用于图像和音频复调音乐转录的多模态变压器
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-01-12 DOI: 10.1016/j.asoc.2026.114643
María Alfaro-Contreras, Noelia Luna-Barahona, Carlos Pérez-Sancho, Jorge Calvo-Zaragoza, Jose J. Valero-Mas
The attainment of structured representations of music sources—commonly referred to as music transcription—has long been a central challenge in the field of Music Information Retrieval. Traditionally, research has addressed this task through Optical Music Recognition for sheet music and through Automatic Music Transcription for audio recordings. More recently, multimodal approaches that integrate both image and audio modalities have emerged, aiming to exploit their complementary strengths. However, these existing methods have so far been evaluated only in controlled settings with monophonic music. In this work, we present the first multimodal image-and-audio framework for polyphonic music transcription, built upon the Transformer architecture. Specifically, we design and evaluate six distinct modality-fusion strategies, differing in the stage at which the modalities are integrated (early, intermediate, or late fusion). Our results demonstrate that multimodality can be beneficial for polyphonic transcription—producing comparable or superior performance across all datasets tested and, in the best-performing experiments, improving the performance of the best unimodal transcription scenario by 9% on average—but its impact depends on the strategy: some fusion schemes yield consistent gains, whereas others fail to improve upon the unimodal baseline.
音乐来源的结构化表示的实现-通常被称为音乐转录-长期以来一直是音乐信息检索领域的核心挑战。传统上,研究通过对乐谱的光学音乐识别和对录音的自动音乐转录来解决这一任务。最近,整合图像和音频模式的多模式方法已经出现,旨在利用它们的互补优势。然而,到目前为止,这些现有的方法仅在单音音乐的受控环境中进行了评估。在这项工作中,我们提出了第一个多模态图像和音频框架,用于复调音乐转录,建立在Transformer架构之上。具体来说,我们设计并评估了六种不同的模式融合策略,这些策略在模式整合的阶段不同(早期、中期或晚期融合)。我们的研究结果表明,多模态对复调转录是有益的——在所有测试的数据集中产生相当或更好的性能,并且在表现最好的实验中,将最佳单模态转录场景的性能平均提高9%——但其影响取决于策略:一些融合方案产生一致的收益,而另一些融合方案在单模态基线上没有改善。
{"title":"Multimodal transformers for image and audio polyphonic music transcription","authors":"María Alfaro-Contreras,&nbsp;Noelia Luna-Barahona,&nbsp;Carlos Pérez-Sancho,&nbsp;Jorge Calvo-Zaragoza,&nbsp;Jose J. Valero-Mas","doi":"10.1016/j.asoc.2026.114643","DOIUrl":"10.1016/j.asoc.2026.114643","url":null,"abstract":"<div><div>The attainment of structured representations of music sources—commonly referred to as music transcription—has long been a central challenge in the field of Music Information Retrieval. Traditionally, research has addressed this task through Optical Music Recognition for sheet music and through Automatic Music Transcription for audio recordings. More recently, multimodal approaches that integrate both image and audio modalities have emerged, aiming to exploit their complementary strengths. However, these existing methods have so far been evaluated only in controlled settings with monophonic music. In this work, we present the first multimodal image-and-audio framework for polyphonic music transcription, built upon the Transformer architecture. Specifically, we design and evaluate six distinct modality-fusion strategies, differing in the stage at which the modalities are integrated (early, intermediate, or late fusion). Our results demonstrate that multimodality can be beneficial for polyphonic transcription—producing comparable or superior performance across all datasets tested and, in the best-performing experiments, improving the performance of the best unimodal transcription scenario by 9% on average—but its impact depends on the strategy: some fusion schemes yield consistent gains, whereas others fail to improve upon the unimodal baseline.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"191 ","pages":"Article 114643"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039722","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}
引用次数: 0
AssemPrompt: A lightweight multi-module collaborative framework for efficient and accurate medical image segmentation AssemPrompt:一个轻量级的多模块协作框架,用于高效准确的医学图像分割
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-02-10 DOI: 10.1016/j.asoc.2026.114801
Hongyan Zhao, Yunteng Hu, Shijie Xu
Medical image segmentation tasks urgently require efficient modeling of multi-level structures and multi-scale boundary features of organs/tissues under limited computational resources. Existing methods struggle to balance high-precision segmentation requirements with real-time constraints for clinical deployment. This paper proposes a multi-module collaborative architecture for medical image segmentation, with its core comprising: (1) The AssemFormer module employs a dual-path parallel mechanism, integrating depthwise separable convolution and linear-complexity self-attention to unify local anatomical details and long-range contextual dependencies through a unified tensor assembly strategy; (2) The PromptGenBlock module introduces a dynamic prompt mechanism based on input distribution, adaptively enhancing target region features using a learnable parameter library while providing category guidance; (3) The GABD module effectively establishes cross-scale spatial associations and strengthens discriminative expression of lesions or boundaries through high-low frequency feature decoupling aggregation, variable dilation rate convolution, and mask-aware attention; (4) The FusionModule combines LHMC multi-scale context modeling with CAM channel attention recalibration, enhancing feature fusion stability and target selectivity through normalization and nonlinear activation. This lightweight architecture achieves a favorable balance between accuracy and efficiency, reaching 81.26% mIoU and 89.66% DSC on ISIC2018 with only 0.0914 M parameters and 0.0897G FLOPs, and shows strong performance across GlaS, ISIC2016, ISIC2017 and CVC-ClinicDB.
医学图像分割任务迫切需要在有限的计算资源下对器官/组织的多层次结构和多尺度边界特征进行高效建模。现有的方法难以平衡高精度分割要求和临床部署的实时限制。本文提出了一种医学图像分割的多模块协同架构,其核心包括:(1)AssemFormer模块采用双路径并行机制,结合深度可分卷积和线性复杂度自关注,通过统一张量组装策略统一局部解剖细节和远程上下文依赖关系;(2) PromptGenBlock模块引入基于输入分布的动态提示机制,在提供类别引导的同时,利用可学习的参数库自适应增强目标区域特征;(3) GABD模块通过高低频特征解耦聚集、变膨胀率卷积和掩模感知关注,有效地建立了跨尺度的空间关联,加强了病灶或边界的判别表达;(4) FusionModule将LHMC多尺度上下文建模与CAM通道注意力再校准相结合,通过归一化和非线性激活增强特征融合稳定性和目标选择性。该轻量级架构在精度和效率之间取得了良好的平衡,仅使用0.0914 M参数和0.0897G FLOPs,在ISIC2018上达到81.26% mIoU和89.66% DSC,并且在GlaS, ISIC2016, ISIC2017和CVC-ClinicDB上表现出强劲的性能。
{"title":"AssemPrompt: A lightweight multi-module collaborative framework for efficient and accurate medical image segmentation","authors":"Hongyan Zhao,&nbsp;Yunteng Hu,&nbsp;Shijie Xu","doi":"10.1016/j.asoc.2026.114801","DOIUrl":"10.1016/j.asoc.2026.114801","url":null,"abstract":"<div><div>Medical image segmentation tasks urgently require efficient modeling of multi-level structures and multi-scale boundary features of organs/tissues under limited computational resources. Existing methods struggle to balance high-precision segmentation requirements with real-time constraints for clinical deployment. This paper proposes a multi-module collaborative architecture for medical image segmentation, with its core comprising: (1) The AssemFormer module employs a dual-path parallel mechanism, integrating depthwise separable convolution and linear-complexity self-attention to unify local anatomical details and long-range contextual dependencies through a unified tensor assembly strategy; (2) The PromptGenBlock module introduces a dynamic prompt mechanism based on input distribution, adaptively enhancing target region features using a learnable parameter library while providing category guidance; (3) The GABD module effectively establishes cross-scale spatial associations and strengthens discriminative expression of lesions or boundaries through high-low frequency feature decoupling aggregation, variable dilation rate convolution, and mask-aware attention; (4) The FusionModule combines LHMC multi-scale context modeling with CAM channel attention recalibration, enhancing feature fusion stability and target selectivity through normalization and nonlinear activation. This lightweight architecture achieves a favorable balance between accuracy and efficiency, reaching 81.26% mIoU and 89.66% DSC on ISIC2018 with only 0.0914 M parameters and 0.0897G FLOPs, and shows strong performance across GlaS, ISIC2016, ISIC2017 and CVC-ClinicDB.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"192 ","pages":"Article 114801"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191851","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}
引用次数: 0
Exploring barriers to metaverse-enabled healthcare 4.0 patient monitoring using interval-valued pythagorean fuzzy DEMATEL 探索使用区间值毕达哥拉斯模糊DEMATEL实现支持元数据的医疗保健4.0患者监控的障碍
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-02-10 DOI: 10.1016/j.asoc.2026.114737
Sachin B. Khot , Shweta S. Patil , Yu-Chen Hu , Juber M. Mulla
The merging of immersive technologies with real-time clinical data in healthcare 4.0 has significantly transformed patient monitoring and management in the metaverse. This study systematically reviewed the applications of metaverse technologies in healthcare. This study further advances the examination of the challenges of adopting metaverse-enabled healthcare systems using an Interval-Valued Pythagorean Fuzzy Decision-Making Trial and Evaluation Laboratory (IVPF-DEMATEL) framework. To evaluate the challenges, this study considered inputs based on evidence received from experts at a prominent healthcare institution in Western Maharashtra, India. A systematic literature review identified 12 critical challenges in the technical, regulatory, social, and economic domains. Expert-based IVPF-DEMATEL evaluations revealed that data security and privacy, infrastructure and interoperability, and regulatory compliance are the most prominent causative barriers. User experience, cultural acceptance, and sustainability concerns have emerged as downstream effects. The cause–and–effect analysis highlights the hierarchical structure of these barriers, demonstrating that addressing core technical and regulatory issues will generate positive ripple effects across user-centric and sociocultural dimensions. The findings provide actionable insights for policymakers, healthcare managers, and technology developers, underscoring the need for secure, interoperable, and ethically governed digital ecosystems to realize the metaverse’s transformative potential in healthcare. To enhance methodological robustness, inter-expert consistency analysis and sensitivity analysis using ±10 % variation in fuzzy membership bounds are conducted. The findings demonstrate that both the cause–and–effect classification and prominence rankings remain stable under moderate perturbations, confirming the reliability of the proposed framework. This study contributes to the growing body of Healthcare 4.0 research by offering empirical evidence from the Indian context and proposing strategic directions for the sustainable and equitable adoption of immersive patient care technologies.
在医疗保健4.0中,沉浸式技术与实时临床数据的融合极大地改变了虚拟世界中的患者监测和管理。本研究系统回顾了元宇宙技术在医疗保健中的应用。本研究使用区间值毕达哥拉模糊决策试验和评估实验室(IVPF-DEMATEL)框架进一步推进了对采用meta -enabled医疗保健系统的挑战的研究。为了评估这些挑战,本研究考虑了来自印度西马哈拉施特拉邦一家著名医疗机构专家的证据。一项系统的文献综述确定了技术、监管、社会和经济领域的12个关键挑战。基于专家的IVPF-DEMATEL评估显示,数据安全和隐私、基础设施和互操作性以及法规遵从性是最突出的导致障碍。用户体验、文化接受度和可持续性问题已成为下游效应。因果分析强调了这些障碍的层次结构,表明解决核心技术和监管问题将在以用户为中心和社会文化层面产生积极的连锁反应。研究结果为政策制定者、医疗保健管理人员和技术开发人员提供了可操作的见解,强调了对安全、可互操作和道德治理的数字生态系统的需求,以实现虚拟世界在医疗保健领域的变革潜力。为了提高方法的稳健性,采用±10 %的模糊隶属度范围变化进行专家间一致性分析和敏感性分析。研究结果表明,在适度扰动下,因果分类和显著性排名都保持稳定,证实了所提出框架的可靠性。本研究通过提供来自印度背景的经验证据,并提出可持续和公平采用沉浸式患者护理技术的战略方向,为不断增长的医疗保健4.0研究做出了贡献。
{"title":"Exploring barriers to metaverse-enabled healthcare 4.0 patient monitoring using interval-valued pythagorean fuzzy DEMATEL","authors":"Sachin B. Khot ,&nbsp;Shweta S. Patil ,&nbsp;Yu-Chen Hu ,&nbsp;Juber M. Mulla","doi":"10.1016/j.asoc.2026.114737","DOIUrl":"10.1016/j.asoc.2026.114737","url":null,"abstract":"<div><div>The merging of immersive technologies with real-time clinical data in healthcare 4.0 has significantly transformed patient monitoring and management in the metaverse. This study systematically reviewed the applications of metaverse technologies in healthcare. This study further advances the examination of the challenges of adopting metaverse-enabled healthcare systems using an Interval-Valued Pythagorean Fuzzy Decision-Making Trial and Evaluation Laboratory (IVPF-DEMATEL) framework. To evaluate the challenges, this study considered inputs based on evidence received from experts at a prominent healthcare institution in Western Maharashtra, India. A systematic literature review identified 12 critical challenges in the technical, regulatory, social, and economic domains. Expert-based IVPF-DEMATEL evaluations revealed that data security and privacy, infrastructure and interoperability, and regulatory compliance are the most prominent causative barriers. User experience, cultural acceptance, and sustainability concerns have emerged as downstream effects. The cause–and–effect analysis highlights the hierarchical structure of these barriers, demonstrating that addressing core technical and regulatory issues will generate positive ripple effects across user-centric and sociocultural dimensions. The findings provide actionable insights for policymakers, healthcare managers, and technology developers, underscoring the need for secure, interoperable, and ethically governed digital ecosystems to realize the metaverse’s transformative potential in healthcare. To enhance methodological robustness, inter-expert consistency analysis and sensitivity analysis using ±10 % variation in fuzzy membership bounds are conducted. The findings demonstrate that both the cause–and–effect classification and prominence rankings remain stable under moderate perturbations, confirming the reliability of the proposed framework. This study contributes to the growing body of Healthcare 4.0 research by offering empirical evidence from the Indian context and proposing strategic directions for the sustainable and equitable adoption of immersive patient care technologies.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"192 ","pages":"Article 114737"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191881","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}
引用次数: 0
Adaptive ant colony optimization with crossover-guided search for dynamic multi-criteria traveling salesman problem 动态多准则旅行商问题的交叉引导搜索自适应蚁群优化
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-01-21 DOI: 10.1016/j.asoc.2026.114668
Sanjai Pathak , Ashish Mani , Amlan Chatterjee
Many real-world instances of the dynamic multi-criteria traveling salesman problem (DMC-TSP) present unpredictable challenges, making them complex dynamic optimization problems. These challenges arise from factors such as the addition of new locations, fluctuating travel times and costs due to changing traffic or weather conditions, or cancellations of pre-scheduled stops, requiring real-time adaptive optimization strategies. While ant colony optimization (ACO) has proven effective for static optimization problems, existing ACO variants have limited adaptability for dynamic multi-criteria environments. To address this significant gap, we introduce a generic test problem generator explicitly for DMC-TSP, capable of creating benchmark problems with known global-optimal solutions. We then propose a novel self-adaptive ant optimization (SAAO) algorithm tailored for DMC-TSP, integrating two new edge assembly crossover operators and a self-adaptive local search operator, explicitly designed to mitigate local-optima stagnation observed in traditional swarm algorithms. Our method demonstrates an effective balance between exploration and exploitation. Comprehensive comparative experiments against four state-of-the-art algorithms, supported by rigorous statistical validation, confirm that the proposed method significantly outperforms existing techniques—achieving a 15–22% improvement in offline performance over baseline algorithms—in terms of adaptability, robustness, and overall efficiency in solving DMC-TSPs.
动态多准则旅行商问题(DMC-TSP)的许多实际实例都存在不可预测的挑战,使其成为复杂的动态优化问题。这些挑战来自于增加新地点、交通或天气条件变化导致的旅行时间和成本波动、或取消预定站点等因素,需要实时自适应优化策略。虽然蚁群算法在静态优化问题上是有效的,但现有的蚁群算法对动态多准则环境的适应性有限。为了解决这一重大差距,我们明确地为DMC-TSP引入了一个通用的测试问题生成器,能够用已知的全局最优解创建基准问题。然后,我们提出了一种针对DMC-TSP的新型自适应蚂蚁优化(SAAO)算法,该算法集成了两个新的边缘组装交叉算子和一个自适应局部搜索算子,明确设计用于减轻传统群算法中观察到的局部最优停滞。我们的方法证明了勘探和开采之间的有效平衡。在严格的统计验证的支持下,与四种最先进的算法进行了全面的比较实验,证实了所提出的方法在求解dmc - tsp的适应性、鲁棒性和整体效率方面显著优于现有技术,比基线算法的离线性能提高了15-22%。
{"title":"Adaptive ant colony optimization with crossover-guided search for dynamic multi-criteria traveling salesman problem","authors":"Sanjai Pathak ,&nbsp;Ashish Mani ,&nbsp;Amlan Chatterjee","doi":"10.1016/j.asoc.2026.114668","DOIUrl":"10.1016/j.asoc.2026.114668","url":null,"abstract":"<div><div>Many real-world instances of the dynamic multi-criteria traveling salesman problem (DMC-TSP) present unpredictable challenges, making them complex dynamic optimization problems. These challenges arise from factors such as the addition of new locations, fluctuating travel times and costs due to changing traffic or weather conditions, or cancellations of pre-scheduled stops, requiring real-time adaptive optimization strategies. While ant colony optimization (ACO) has proven effective for static optimization problems, existing ACO variants have limited adaptability for dynamic multi-criteria environments. To address this significant gap, we introduce a generic test problem generator explicitly for DMC-TSP, capable of creating benchmark problems with known global-optimal solutions. We then propose a novel self-adaptive ant optimization (SAAO) algorithm tailored for DMC-TSP, integrating two new edge assembly crossover operators and a self-adaptive local search operator, explicitly designed to mitigate local-optima stagnation observed in traditional swarm algorithms. Our method demonstrates an effective balance between exploration and exploitation. Comprehensive comparative experiments against four state-of-the-art algorithms, supported by rigorous statistical validation, confirm that the proposed method significantly outperforms existing techniques—achieving a 15–22% improvement in offline performance over baseline algorithms—in terms of adaptability, robustness, and overall efficiency in solving DMC-TSPs.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"191 ","pages":"Article 114668"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080110","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}
引用次数: 0
Multi-objective evolutionary algorithms and integer programming for the optimization of underwater acoustic sensor network design 基于多目标进化算法和整数规划的水声传感器网络优化设计
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-01-21 DOI: 10.1016/j.asoc.2026.114685
Laurent Lemarchand , Ronan Serré , Bilal Latrach , Mathis Hamelotte , Catherine Dezan , David Dellong , Myriam Lajaunie
Acoustic surveillance of oceans is of great interest for monitoring both wildlife and anthropogenic activities. In this paper, we focus on passive, non-communicating sonar systems that are suitable for efficiently monitoring the marine environment. With such a network, no live tracking is possible; instead, we aim to record activities in the most discreet manner for surveillance purposes. General guidelines can be established to assist in designing a network of underwater acoustic sensors. These guidelines serve as both objectives and constraints, such as sensor detection capabilities, deployment area alternatives, acoustic propagation properties, and potential noise source areas.
The optimization of network design, in terms of the distribution and locations of sensors (or groups of sensors physically stacked on a mooring line, for example, and referred to as anchors), involves trade-offs between network cost, coverage of the targeted marine area, network redundancy, and localization capabilities. In this paper, we propose several approaches to address these various constraints and optimization objectives. These approaches rely on numerical simulations of underwater acoustic propagation, which allow for accounting for environmental characteristics such as bathymetry, sea floor properties, and sound speed profiles.
A discretized model of the network design problem is presented. It relies on a grid representation of anchor locations, and a complete Integer Programming formulation of the problem is defined. The model is used to identify sets of solutions that represent trade-offs among the different optimization criteria. Both exact and heuristic multi-objective approaches are proposed to exploit the model. The former are applicable to smaller sets of instances and optimization objectives than the latter. However, the exact approaches enable the derivation of precise bounds on the optimized criteria, as well as initial solutions that can enhance heuristic multi-objective methods based on evolutionary multi-objective optimization frameworks. This supports the designer’s choice of the final network configuration.
Numerical experiments conducted on a set of sixteen semi-synthetic test cases, as well as on an actual network deployed at sea, demonstrate that an exact ϵ-constraint method is applicable to industrial-scale instances involving up to 2000 monitoring points. It provides the complete Pareto set for the coverage vs. cost problem in a few minutes. One of the proposed heuristic methods yields, in less than a second, sets of solutions with a loss of quality below 1% for the same design problem. When the triangulation capabilities of the network are optimized as a complementary objective for noise source localization, hybrid heuristic approaches are able to improve the results of the basic algorithm by up to 156% on average.
海洋声学监测对于监测野生动物和人为活动都具有重要意义。在本文中,我们重点研究了适用于有效监测海洋环境的被动、非通信声纳系统。有了这样的网络,就不可能进行实时跟踪;相反,我们的目标是以最谨慎的方式记录活动,用于监视目的。可以建立一般准则来协助设计水声传感器网络。这些指导方针既是目标也是约束条件,例如传感器探测能力、部署区域选择、声学传播特性和潜在噪声源区域。根据传感器的分布和位置(或物理上堆叠在系泊线上的传感器组,例如,称为锚),网络设计的优化涉及网络成本、目标海洋区域的覆盖范围、网络冗余和定位能力之间的权衡。在本文中,我们提出了几种方法来解决这些不同的约束和优化目标。这些方法依赖于水声传播的数值模拟,从而考虑到环境特征,如测深、海底特性和声速分布。提出了网络设计问题的离散化模型。它依赖于锚点位置的网格表示,并定义了问题的完整整数规划公式。该模型用于识别代表不同优化标准之间权衡的解决方案集。提出了精确多目标和启发式多目标两种方法来开发该模型。前者比后者适用于更小的实例集和优化目标。然而,精确的方法可以推导出优化准则的精确边界,以及可以增强基于进化多目标优化框架的启发式多目标方法的初始解。这支持设计者选择最终的网络配置。在16个半综合测试用例和海上部署的实际网络上进行的数值实验表明,一种精确的ϵ-constraint方法适用于涉及多达2000个监测点的工业规模实例。它在几分钟内提供了覆盖与成本问题的完整Pareto集。其中一种提出的启发式方法,在不到一秒钟的时间内,为相同的设计问题提供了一组质量损失低于1%的解决方案。当将网络的三角测量能力作为噪声源定位的补充目标进行优化时,混合启发式方法平均可将基本算法的结果提高156%。
{"title":"Multi-objective evolutionary algorithms and integer programming for the optimization of underwater acoustic sensor network design","authors":"Laurent Lemarchand ,&nbsp;Ronan Serré ,&nbsp;Bilal Latrach ,&nbsp;Mathis Hamelotte ,&nbsp;Catherine Dezan ,&nbsp;David Dellong ,&nbsp;Myriam Lajaunie","doi":"10.1016/j.asoc.2026.114685","DOIUrl":"10.1016/j.asoc.2026.114685","url":null,"abstract":"<div><div>Acoustic surveillance of oceans is of great interest for monitoring both wildlife and anthropogenic activities. In this paper, we focus on passive, non-communicating sonar systems that are suitable for efficiently monitoring the marine environment. With such a network, no live tracking is possible; instead, we aim to record activities in the most discreet manner for surveillance purposes. General guidelines can be established to assist in designing a network of underwater acoustic sensors. These guidelines serve as both objectives and constraints, such as sensor detection capabilities, deployment area alternatives, acoustic propagation properties, and potential noise source areas.</div><div>The optimization of network design, in terms of the distribution and locations of sensors (or groups of sensors physically stacked on a mooring line, for example, and referred to as <em>anchors</em>), involves trade-offs between network cost, coverage of the targeted marine area, network redundancy, and localization capabilities. In this paper, we propose several approaches to address these various constraints and optimization objectives. These approaches rely on numerical simulations of underwater acoustic propagation, which allow for accounting for environmental characteristics such as bathymetry, sea floor properties, and sound speed profiles.</div><div>A discretized model of the network design problem is presented. It relies on a grid representation of anchor locations, and a complete Integer Programming formulation of the problem is defined. The model is used to identify sets of solutions that represent trade-offs among the different optimization criteria. Both exact and heuristic multi-objective approaches are proposed to exploit the model. The former are applicable to smaller sets of instances and optimization objectives than the latter. However, the exact approaches enable the derivation of precise bounds on the optimized criteria, as well as initial solutions that can enhance heuristic multi-objective methods based on evolutionary multi-objective optimization frameworks. This supports the designer’s choice of the final network configuration.</div><div>Numerical experiments conducted on a set of sixteen semi-synthetic test cases, as well as on an actual network deployed at sea, demonstrate that an exact <span><math><mi>ϵ</mi></math></span>-constraint method is applicable to industrial-scale instances involving up to 2000 monitoring points. It provides the complete Pareto set for the coverage <em>vs.</em> cost problem in a few minutes. One of the proposed heuristic methods yields, in less than a second, sets of solutions with a loss of quality below 1% for the same design problem. When the triangulation capabilities of the network are optimized as a complementary objective for noise source localization, hybrid heuristic approaches are able to improve the results of the basic algorithm by up to 156% on average.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"191 ","pages":"Article 114685"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080201","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}
引用次数: 0
Reinforcement learning-driven nonsingular fast terminal sliding mode control for pipe crack sealing manipulator 管道缝封机械手的强化学习驱动非奇异快速终端滑模控制
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-01-24 DOI: 10.1016/j.asoc.2026.114708
Santosh Kumar, S.K. Dwivedy
This study explores control challenges associated with the Pipe Crack Sealing Manipulator (PCSM), a tree-shaped robot designed to seal cracks within concrete pipes. Equipped with five specialized branches, the PCSM adeptly navigates both vertical and horizontal pipe environments. A detailed CAD model was developed in SolidWorks, and data were meticulously exported to Simulink for refined control operations. SolidWorks simulations revealed that the PCSM’s fifth branch excelled in extensive crack repairs. The control of pipe crack sealing manipulators presents significant challenges due to uncertain dynamics, random disturbances, signal delay, and sensor noise inherent in confined environments. To overcome these issues, this study proposes a reinforcement learning–driven nonsingular fast terminal sliding mode control (RLNFTSMC) framework that enhances robustness, precision, and adaptability for crack sealing operations. The implemented RLNFTSMC strategy, includes an equivalent part utilizing actor-critic neural networks for dynamic estimation and a switching part that treats the mass matrix as a lumped value. Comparative analysis with nonsingular fast terminal sliding mode control (NFTSMC), adaptive NFTSMC, radial basis function neural network control (RBFNNC), and neural network-enhanced nonsingular fast terminal sliding mode control (NENFTSMC) shows that RLNFTSMC significantly improves the PCSM’s reliability and effectiveness in environments marked by random disturbances and unknown dynamics.
本研究探讨了与管道裂缝密封机械手(PCSM)相关的控制挑战,PCSM是一种设计用于密封混凝土管道裂缝的树形机器人。PCSM配备了五个专门的分支机构,可以熟练地导航垂直和水平管道环境。在SolidWorks中开发了详细的CAD模型,并将数据精心导出到Simulink中进行精细控制操作。SolidWorks模拟显示,PCSM的第五分支机构在广泛的裂缝修复方面表现出色。由于不确定动力学、随机干扰、信号延迟和受限环境中固有的传感器噪声,管道裂缝密封机械手的控制面临着重大挑战。为了克服这些问题,本研究提出了一种强化学习驱动的非奇异快速终端滑模控制(RLNFTSMC)框架,该框架增强了裂缝密封操作的鲁棒性、精度和适应性。实现的RLNFTSMC策略包括一个等效部分,该部分利用行动者批评神经网络进行动态估计,以及一个将质量矩阵视为集总值的切换部分。与非奇异快速终端滑模控制(NFTSMC)、自适应非奇异终端滑模控制(NFTSMC)、径向基函数神经网络控制(RBFNNC)和神经网络增强非奇异快速终端滑模控制(NENFTSMC)的对比分析表明,RLNFTSMC显著提高了PCSM在随机干扰和未知动态环境中的可靠性和有效性。
{"title":"Reinforcement learning-driven nonsingular fast terminal sliding mode control for pipe crack sealing manipulator","authors":"Santosh Kumar,&nbsp;S.K. Dwivedy","doi":"10.1016/j.asoc.2026.114708","DOIUrl":"10.1016/j.asoc.2026.114708","url":null,"abstract":"<div><div>This study explores control challenges associated with the Pipe Crack Sealing Manipulator (PCSM), a tree-shaped robot designed to seal cracks within concrete pipes. Equipped with five specialized branches, the PCSM adeptly navigates both vertical and horizontal pipe environments. A detailed CAD model was developed in SolidWorks, and data were meticulously exported to Simulink for refined control operations. SolidWorks simulations revealed that the PCSM’s fifth branch excelled in extensive crack repairs. The control of pipe crack sealing manipulators presents significant challenges due to uncertain dynamics, random disturbances, signal delay, and sensor noise inherent in confined environments. To overcome these issues, this study proposes a reinforcement learning–driven nonsingular fast terminal sliding mode control (RLNFTSMC) framework that enhances robustness, precision, and adaptability for crack sealing operations. The implemented RLNFTSMC strategy, includes an equivalent part utilizing actor-critic neural networks for dynamic estimation and a switching part that treats the mass matrix as a lumped value. Comparative analysis with nonsingular fast terminal sliding mode control (NFTSMC), adaptive NFTSMC, radial basis function neural network control (RBFNNC), and neural network-enhanced nonsingular fast terminal sliding mode control (NENFTSMC) shows that RLNFTSMC significantly improves the PCSM’s reliability and effectiveness in environments marked by random disturbances and unknown dynamics.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"191 ","pages":"Article 114708"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080293","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}
引用次数: 0
Hybrid-conditional generative adversarial network framework for climate fault detection in vertical farming environments 垂直农业环境气候故障检测的混合条件生成对抗网络框架
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-01-27 DOI: 10.1016/j.asoc.2026.114727
P.K.S. Tejes , Abhishek Dasore , Norhashila Hashim , Bukke Kiran Naik , Challa Monisha Reddy
Fault diagnosis in vertical farming systems (VFS) is constrained by the lack of labeled fault data, especially under rare or complex airflow disturbance conditions. This limits the training and evaluation of machine learning (ML) models for real-time microclimate monitoring and control. This study aims to develop and assess a Hybrid Conditional Generative Adversarial Network (Hybrid-CGAN) to generate synthetic microclimate sequences for specific airflow fault scenarios using computational fluid dynamics (CFD)-based simulation data. Four airflow fault scenarios, such as, reduced airflow from the upper inlet plenum, rack short-circuit, non-functional rear exhaust fan, and high static pressure at the return-duct junction, were simulated using a steady-state CFD model of a multi-tier plant factory. Diurnal perturbations generated 24-hour temporal sequences. Synthetic datasets trained Hybrid-CGAN and were compared with baselines such as Synthetic Minority Over-sampling Technique (SMOTE), Random Oversampling (ROS), and unconditional GANs. Performance was evaluated using Dynamic Time Warping (DTW), Uniform Manifold Approximation and Projection (UMAP) visualizations, and accuracy of Gated Recurrent Unit (GRU), Light Gradient Boosting Machine (LGBM), and k-Nearest Neighbors (k-NN) models. Hybrid-CGAN achieved the lowest DTW score of 0.10 at 100 % CFD data, representing a 61.03 % improvement over standard GAN (DTW = 10.45) and 59.76 % over simulation-only baseline (DTW = 10.12). Classifiers trained on Hybrid-CGAN-augmented datasets achieved accuracies of 0.95 (GRU), 0.96 (LGBM), and 0.94 (k-NN). UMAP showed high clustering similarity between real and synthetic sequences. This study establishes a practical paradigm by integrating CFD-based airflow simulations with Hybrid-CGAN framework for fault-specific synthetic data generation in VFS.
垂直农业系统(VFS)的故障诊断受到缺乏标记故障数据的限制,特别是在罕见或复杂的气流扰动条件下。这限制了用于实时微气候监测和控制的机器学习(ML)模型的训练和评估。本研究旨在开发和评估混合条件生成对抗网络(Hybrid- cgan),利用基于计算流体动力学(CFD)的模拟数据为特定的气流故障场景生成合成的微气候序列。采用多层厂房的稳态CFD模型,模拟了上进气室气流减少、机架短路、后排风机失效、回风管连接处静压过高等4种气流故障情况。日扰动产生24小时时间序列。合成数据集训练Hybrid-CGAN,并与合成少数过采样技术(SMOTE)、随机过采样(ROS)和无条件gan等基线进行比较。使用动态时间扭曲(DTW)、均匀流形逼近和投影(UMAP)可视化以及门通循环单元(GRU)、光梯度增强机(LGBM)和k-近邻(k-NN)模型的精度来评估性能。Hybrid-CGAN在100 % CFD数据下的DTW得分最低,为0.10,比标准GAN (DTW = 10.45)提高了61.03 %,比仅模拟的GAN (DTW = 10.12)提高了59.76 %。在hybrid - cgan增强数据集上训练的分类器的准确率分别为0.95 (GRU)、0.96 (LGBM)和0.94 (k-NN)。UMAP显示真实序列与合成序列具有较高的聚类相似性。该研究通过将基于cfd的气流模拟与Hybrid-CGAN框架相结合,建立了一个实用的范例,用于VFS中特定故障的合成数据生成。
{"title":"Hybrid-conditional generative adversarial network framework for climate fault detection in vertical farming environments","authors":"P.K.S. Tejes ,&nbsp;Abhishek Dasore ,&nbsp;Norhashila Hashim ,&nbsp;Bukke Kiran Naik ,&nbsp;Challa Monisha Reddy","doi":"10.1016/j.asoc.2026.114727","DOIUrl":"10.1016/j.asoc.2026.114727","url":null,"abstract":"<div><div>Fault diagnosis in vertical farming systems (VFS) is constrained by the lack of labeled fault data, especially under rare or complex airflow disturbance conditions. This limits the training and evaluation of machine learning (ML) models for real-time microclimate monitoring and control. This study aims to develop and assess a Hybrid Conditional Generative Adversarial Network (Hybrid-CGAN) to generate synthetic microclimate sequences for specific airflow fault scenarios using computational fluid dynamics (CFD)-based simulation data. Four airflow fault scenarios, such as, reduced airflow from the upper inlet plenum, rack short-circuit, non-functional rear exhaust fan, and high static pressure at the return-duct junction, were simulated using a steady-state CFD model of a multi-tier plant factory. Diurnal perturbations generated 24-hour temporal sequences. Synthetic datasets trained Hybrid-CGAN and were compared with baselines such as Synthetic Minority Over-sampling Technique (SMOTE), Random Oversampling (ROS), and unconditional GANs. Performance was evaluated using Dynamic Time Warping (DTW), Uniform Manifold Approximation and Projection (UMAP) visualizations, and accuracy of Gated Recurrent Unit (GRU), Light Gradient Boosting Machine (LGBM), and k-Nearest Neighbors (k-NN) models. Hybrid-CGAN achieved the lowest DTW score of 0.10 at 100 % CFD data, representing a 61.03 % improvement over standard GAN (DTW = 10.45) and 59.76 % over simulation-only baseline (DTW = 10.12). Classifiers trained on Hybrid-CGAN-augmented datasets achieved accuracies of 0.95 (GRU), 0.96 (LGBM), and 0.94 (k-NN). UMAP showed high clustering similarity between real and synthetic sequences. This study establishes a practical paradigm by integrating CFD-based airflow simulations with Hybrid-CGAN framework for fault-specific synthetic data generation in VFS.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"191 ","pages":"Article 114727"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080194","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}
引用次数: 0
Adaptive multi-region multi-directional competitive swarm optimizer algorithm for large-scale multi-objective problems 大规模多目标问题的自适应多区域多向竞争群优化算法
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-02-05 DOI: 10.1016/j.asoc.2026.114741
Debao Chen , Yuanyuan Ge , Feng Zou , Fangzhen Ge , Lei Zhang
Competitive swarm optimization has shown impressive performance in addressing large-scale multi-objective optimization problems. However, its competitive mechanism can lead to contention among local optimal solutions within the current population. As the decision space grows more complex, the diversity of the population in competitive swarm optimization often decreases, resulting in “premature convergence”. To address this issue, we propose an adaptive multi-region multi-directional competitive swarm optimization algorithm, named AMRCSO. First, this algorithm introduces an adaptive reference-guided subregion division strategy that segments the objective space based on the population’s distribution, enhancing search diversity and coverage. Additionally, a multi-directional exploration strategy is employed to explore each subregion of the decision space, enabling parallel and multi-directional exploration that accelerates convergence to the global optimal solution. Furthermore, an effective population initialization and tri-stage updating strategy are implemented to improve population diversity and enhance convergence accuracy, respectively. Experimental results on large-scale multi-objective benchmark functions, ranging from 500 to 5000 decision variables, demonstrate that the AMRCSO algorithm achieves superior convergence accuracy and distribution compared to existing algorithms for large-scale multi-objective optimization problems.
竞争群算法在解决大规模多目标优化问题方面表现出令人印象深刻的性能。然而,它的竞争机制会导致当前种群中局部最优解之间的竞争。当决策空间变得越来越复杂时,竞争群体优化中的种群多样性往往会降低,从而导致“过早收敛”。为了解决这一问题,我们提出了一种自适应多区域多向竞争群体优化算法AMRCSO。首先,该算法引入自适应参考引导子区域划分策略,根据人口分布对目标空间进行细分,增强搜索多样性和覆盖范围;此外,采用多向探索策略对决策空间的每个子区域进行探索,实现并行和多向探索,加速收敛到全局最优解。采用有效的种群初始化策略和三阶段更新策略,分别提高种群多样性和收敛精度。在500 ~ 5000个决策变量的大规模多目标基准函数上的实验结果表明,与现有算法相比,AMRCSO算法在大规模多目标优化问题上具有更好的收敛精度和分布。
{"title":"Adaptive multi-region multi-directional competitive swarm optimizer algorithm for large-scale multi-objective problems","authors":"Debao Chen ,&nbsp;Yuanyuan Ge ,&nbsp;Feng Zou ,&nbsp;Fangzhen Ge ,&nbsp;Lei Zhang","doi":"10.1016/j.asoc.2026.114741","DOIUrl":"10.1016/j.asoc.2026.114741","url":null,"abstract":"<div><div>Competitive swarm optimization has shown impressive performance in addressing large-scale multi-objective optimization problems. However, its competitive mechanism can lead to contention among local optimal solutions within the current population. As the decision space grows more complex, the diversity of the population in competitive swarm optimization often decreases, resulting in “premature convergence”. To address this issue, we propose an adaptive multi-region multi-directional competitive swarm optimization algorithm, named AMRCSO. First, this algorithm introduces an adaptive reference-guided subregion division strategy that segments the objective space based on the population’s distribution, enhancing search diversity and coverage. Additionally, a multi-directional exploration strategy is employed to explore each subregion of the decision space, enabling parallel and multi-directional exploration that accelerates convergence to the global optimal solution. Furthermore, an effective population initialization and tri-stage updating strategy are implemented to improve population diversity and enhance convergence accuracy, respectively. Experimental results on large-scale multi-objective benchmark functions, ranging from 500 to 5000 decision variables, demonstrate that the AMRCSO algorithm achieves superior convergence accuracy and distribution compared to existing algorithms for large-scale multi-objective optimization problems.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"192 ","pages":"Article 114741"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190706","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}
引用次数: 0
ProtoMark: A transfer prototype network with Markov-enhanced features for few-shot intrusion detection ProtoMark:一个传输原型网络,具有马尔可夫增强功能,用于少量入侵检测
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-02-10 DOI: 10.1016/j.asoc.2026.114806
Yixin Xu , Jingwen Wang , Yue Chen , Junfeng Wang , He Sun , Zhiyang Fang
With the increasing prevalence of targeted attacks against Local Area Networks (LANs), the demand for adaptive intrusion detection tailored to heterogeneous networks has become increasingly prominent. However, as penetration data from LANs is typically scarce due to its sensitivity, conventional data-driven approaches face significant limitations. To fill this gap, we propose ProtoMark, a highly transferable prototype network that incorporates Markovian traffic characteristics into feature representation, establishing an enhanced Transfer Prototype Network framework for effective intrusion detection. Specifically, the model fuses Markov features of traffic byte sequences with statistical features to generate new feature images, thereby enriching the learnable content of the model. ProtoMark combines fully connected neural networks and prototype networks to further enhance the model’s generalization ability to build a transferable prototype network model. The model achieves intrusion detection by constructing prototypes of different types of traffic and optimizing its transfer performance by fine-tuning the fully connected layer (FC) parameters. Experimental results show that ProtoMark performs superior network intrusion detection on the CICIDS2017 and CICDDoS2019 datasets, outperforming existing state-of-the-art methods. Under the 2-way 5-shot setting, the average accuracy of ProtoMark is 94.76% and 95.37% on the two datasets. Additionally, ablation studies further validate the effectiveness of Markov features and the significant contribution of the transfer layer to the model’s generalization ability.
随着针对局域网(lan)的针对性攻击日益普遍,针对异构网络的自适应入侵检测需求日益突出。然而,由于来自局域网的渗透数据由于其敏感性通常是稀缺的,传统的数据驱动方法面临着很大的局限性。为了填补这一空白,我们提出了ProtoMark,这是一个高度可转移的原型网络,它将马尔可夫流量特征融入到特征表示中,建立了一个增强的传输原型网络框架,用于有效的入侵检测。具体来说,该模型将交通字节序列的马尔可夫特征与统计特征融合,生成新的特征图像,从而丰富了模型的可学习内容。ProtoMark将全连接神经网络与原型网络相结合,进一步增强模型的泛化能力,构建可转移的原型网络模型。该模型通过构造不同类型流量的原型,并通过微调FC (fully connected layer)参数来优化其传输性能,从而实现入侵检测。实验结果表明,ProtoMark在CICIDS2017和CICDDoS2019数据集上的网络入侵检测性能优于现有的最先进方法。在2路5枪设置下,ProtoMark在两个数据集上的平均准确率分别为94.76%和95.37%。此外,消融研究进一步验证了马尔可夫特征的有效性以及传递层对模型泛化能力的重要贡献。
{"title":"ProtoMark: A transfer prototype network with Markov-enhanced features for few-shot intrusion detection","authors":"Yixin Xu ,&nbsp;Jingwen Wang ,&nbsp;Yue Chen ,&nbsp;Junfeng Wang ,&nbsp;He Sun ,&nbsp;Zhiyang Fang","doi":"10.1016/j.asoc.2026.114806","DOIUrl":"10.1016/j.asoc.2026.114806","url":null,"abstract":"<div><div>With the increasing prevalence of targeted attacks against Local Area Networks (LANs), the demand for adaptive intrusion detection tailored to heterogeneous networks has become increasingly prominent. However, as penetration data from LANs is typically scarce due to its sensitivity, conventional data-driven approaches face significant limitations. To fill this gap, we propose ProtoMark, a highly transferable prototype network that incorporates Markovian traffic characteristics into feature representation, establishing an enhanced Transfer Prototype Network framework for effective intrusion detection. Specifically, the model fuses Markov features of traffic byte sequences with statistical features to generate new feature images, thereby enriching the learnable content of the model. ProtoMark combines fully connected neural networks and prototype networks to further enhance the model’s generalization ability to build a transferable prototype network model. The model achieves intrusion detection by constructing prototypes of different types of traffic and optimizing its transfer performance by fine-tuning the fully connected layer (FC) parameters. Experimental results show that ProtoMark performs superior network intrusion detection on the CICIDS2017 and CICDDoS2019 datasets, outperforming existing state-of-the-art methods. Under the 2-way 5-shot setting, the average accuracy of ProtoMark is 94.76% and 95.37% on the two datasets. Additionally, ablation studies further validate the effectiveness of Markov features and the significant contribution of the transfer layer to the model’s generalization ability.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"192 ","pages":"Article 114806"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191884","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}
引用次数: 0
Topic-aware influence maximization based on deep reinforcement learning and group relative optimization 基于深度强化学习和群体相对优化的主题感知影响最大化
IF 6.6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Pub Date : 2026-04-01 Epub Date: 2026-02-03 DOI: 10.1016/j.asoc.2026.114786
Yingqi Zou , Guanyu Li , Yanhao Wang , Bo Ning , Shaohan Wang
Information diffusion on social networks is increasingly complex and diverse. Identifying suitable users to recommend and spread specific information is critical for social network analysis. The Topic-aware Influence Maximization (TIM) problem aims to identify a seed set that maximizes the influence spread under a given topic distribution. However, existing TIM methods suffer from severe computational inefficiencies. Meanwhile, current Deep Reinforcement Learning (DRL)-based methods mostly ignore the interplay between network structure and topic heterogeneity. To address these challenges, this paper proposes GR-TIM, an end-to-end DRL-based framework for TIM. GR-TIM first estimates the pre-global influence using Graph Neural Networks (GNNs). Then, a group relative optimization strategy partitions users based on topic and community structures. We further leverage intra-group collaboration to apply the global-local optimization paradigm to agent training and inter-group competition to achieve adaptive seed selection for the target topic. Experiments on six real-world datasets demonstrate that GR-TIM outperforms state-of-the-art DRL-based methods in terms of multi-topic influence spread and reduces runtime by two to three orders of magnitude compared to existing simulation-based methods.
社交网络上的信息传播越来越复杂和多样化。识别合适的用户来推荐和传播特定的信息是社会网络分析的关键。主题感知影响最大化(TIM)问题旨在识别给定主题分布下影响传播最大化的种子集。然而,现有的TIM方法存在严重的计算效率低下的问题。同时,目前基于深度强化学习(Deep Reinforcement Learning, DRL)的方法大多忽略了网络结构与主题异质性之间的相互作用。为了应对这些挑战,本文提出了基于drl的端到端TIM框架GR-TIM。GR-TIM首先使用图神经网络(gnn)估计前全球影响。然后,采用基于主题和社区结构的群体相对优化策略对用户进行划分。我们进一步利用组内协作,将全局-局部优化范式应用于智能体训练和组间竞争,以实现目标主题的自适应种子选择。在六个真实数据集上的实验表明,GR-TIM在多主题影响传播方面优于最先进的基于drl的方法,并且与现有的基于仿真的方法相比,运行时间缩短了两到三个数量级。
{"title":"Topic-aware influence maximization based on deep reinforcement learning and group relative optimization","authors":"Yingqi Zou ,&nbsp;Guanyu Li ,&nbsp;Yanhao Wang ,&nbsp;Bo Ning ,&nbsp;Shaohan Wang","doi":"10.1016/j.asoc.2026.114786","DOIUrl":"10.1016/j.asoc.2026.114786","url":null,"abstract":"<div><div>Information diffusion on social networks is increasingly complex and diverse. Identifying suitable users to recommend and spread specific information is critical for social network analysis. The Topic-aware Influence Maximization (TIM) problem aims to identify a seed set that maximizes the influence spread under a given topic distribution. However, existing TIM methods suffer from severe computational inefficiencies. Meanwhile, current Deep Reinforcement Learning (DRL)-based methods mostly ignore the interplay between network structure and topic heterogeneity. To address these challenges, this paper proposes GR-TIM, an end-to-end DRL-based framework for TIM. GR-TIM first estimates the pre-global influence using Graph Neural Networks (GNNs). Then, a group relative optimization strategy partitions users based on topic and community structures. We further leverage intra-group collaboration to apply the global-local optimization paradigm to agent training and inter-group competition to achieve adaptive seed selection for the target topic. Experiments on six real-world datasets demonstrate that GR-TIM outperforms state-of-the-art DRL-based methods in terms of multi-topic influence spread and reduces runtime by two to three orders of magnitude compared to existing simulation-based methods.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"192 ","pages":"Article 114786"},"PeriodicalIF":6.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191834","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}
引用次数: 0
期刊
Applied Soft Computing
全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.
﹀
mjeditor
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
Book学术官方微信
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1