Neurocomputing最新文献

Enhancing shape bias for object detection 增强物体检测的形状偏差

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-06 DOI: 10.1016/j.neucom.2026.132931

Jiwen Tang , Gu Wang , Ruida Zhang , Xiangyang Ji

Convolutional Neural Networks (CNNs) are widely used for object detection tasks, whereas recent studies have shown that they rely more on texture rather than shape for object recognition, a phenomenon known as texture bias. This bias makes them vulnerable to image corruptions, domain shifts, and adversarial perturbations, posing significant challenges for real-world deployment, especially in safety-critical and industrial applications. Despite its significance, texture bias in object detection remains largely underexplored. To address this gap, we first conduct a comprehensive analysis of texture bias across multiple widely-used CNN-based detection architectures, demonstrating the widespread presence and detrimental impact of this issue. Motivated by these findings, we propose a simple yet effective method, TexDrop, to increase shape bias in CNNs and therefore improve their accuracy and robustness. Specifically, TexDrop randomly drops out the texture and color of the training images through straightforward edge detection, forcing models to learn to detect objects based on their shape, thus increasing shape bias. Unlike prior approaches that require architectural modifications, extensive additional training data or complex regularization schemes, TexDrop is model-agnostic, easy to integrate into existing training pipelines, and incurs negligible computational overhead. Intensive experiments on Pascal VOC, COCO, and various corrupted COCO datasets demonstrate that TexDrop not only improves detection performance across multiple architectures but also consistently enhances robustness against various image corruptions and texture variations. Our study provides empirical insights into texture dependence in object detectors and contributes a practical solution for developing more robust and reliable object detection systems in real-world applications.

卷积神经网络（cnn）被广泛用于目标检测任务，然而最近的研究表明，它们更多地依赖于纹理而不是形状来识别目标，这种现象被称为纹理偏差。这种偏差使它们容易受到图像损坏、域移位和对抗性扰动的影响，对现实世界的部署构成了重大挑战，特别是在安全关键和工业应用中。尽管具有重要意义，但纹理偏差在目标检测中的应用仍未得到充分的研究。为了解决这一差距，我们首先对多种广泛使用的基于cnn的检测架构进行了纹理偏差的综合分析，证明了该问题的广泛存在和有害影响。基于这些发现，我们提出了一种简单而有效的方法TexDrop来增加cnn的形状偏差，从而提高其准确性和鲁棒性。具体来说，TexDrop通过直接的边缘检测随机掉去训练图像的纹理和颜色，迫使模型学习根据物体的形状来检测物体，从而增加形状偏差。与之前需要修改架构、大量额外训练数据或复杂正则化方案的方法不同，TexDrop与模型无关，易于集成到现有的训练管道中，并且产生的计算开销可以忽略不计。在Pascal VOC、COCO和各种损坏的COCO数据集上进行的大量实验表明，TexDrop不仅提高了跨多个架构的检测性能，而且始终增强了对各种图像损坏和纹理变化的鲁棒性。我们的研究为物体检测器的纹理依赖性提供了经验见解，并为在实际应用中开发更健壮和可靠的物体检测系统提供了实用的解决方案。

{"title":"Enhancing shape bias for object detection","authors":"Jiwen Tang , Gu Wang , Ruida Zhang , Xiangyang Ji","doi":"10.1016/j.neucom.2026.132931","DOIUrl":"10.1016/j.neucom.2026.132931","url":null,"abstract":"<div><div>Convolutional Neural Networks (CNNs) are widely used for object detection tasks, whereas recent studies have shown that they rely more on texture rather than shape for object recognition, a phenomenon known as texture bias. This bias makes them vulnerable to image corruptions, domain shifts, and adversarial perturbations, posing significant challenges for real-world deployment, especially in safety-critical and industrial applications. Despite its significance, texture bias in object detection remains largely underexplored. To address this gap, we first conduct a comprehensive analysis of texture bias across multiple widely-used CNN-based detection architectures, demonstrating the widespread presence and detrimental impact of this issue. Motivated by these findings, we propose a simple yet effective method, TexDrop, to increase shape bias in CNNs and therefore improve their accuracy and robustness. Specifically, TexDrop randomly drops out the texture and color of the training images through straightforward edge detection, forcing models to learn to detect objects based on their shape, thus increasing shape bias. Unlike prior approaches that require architectural modifications, extensive additional training data or complex regularization schemes, TexDrop is model-agnostic, easy to integrate into existing training pipelines, and incurs negligible computational overhead. Intensive experiments on Pascal VOC, COCO, and various corrupted COCO datasets demonstrate that TexDrop not only improves detection performance across multiple architectures but also consistently enhances robustness against various image corruptions and texture variations. Our study provides empirical insights into texture dependence in object detectors and contributes a practical solution for developing more robust and reliable object detection systems in real-world applications.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132931"},"PeriodicalIF":6.5,"publicationDate":"2026-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

TabNSA: Native sparse attention for efficient tabular data learning TabNSA：用于高效表格数据学习的原生稀疏关注

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132928

Ali Eslamian , Qiang Cheng

Tabular data poses unique challenges for deep learning due to its heterogeneous feature types, lack of spatial structure, and often limited sample sizes. We propose TabNSA, a novel deep learning framework that integrates Native Sparse Attention (NSA) with a TabMixer backbone to efficiently model tabular data. TabNSA tackles computational and representational challenges by dynamically focusing on relevant feature subsets per instance. The NSA module employs a hierarchical sparse attention mechanism, including token compression, selective preservation, and localized sliding windows, to significantly reduce the quadratic complexity of standard attention operations while addressing feature heterogeneity. Complementing this, the TabMixer backbone captures complex, non-linear dependencies through parallel multilayer perceptron (MLP) branches with independent parameters. These modules are synergistically combined via element-wise summation and mean pooling, enabling TabNSA to model both global context and fine-grained interactions. Extensive experiments across supervised and transfer learning settings show that TabNSA consistently outperforms state-of-the-art deep learning models. Furthermore, by augmenting TabNSA with a fine-tuned large language model (LLM), we enable it to effectively address Few-Shot Learning challenges through language-guided generalization on diverse tabular benchmarks. Code available on: https://github.com/aseslamian/TabNSA.

表格数据由于其异构的特征类型、缺乏空间结构和通常有限的样本量，给深度学习带来了独特的挑战。我们提出了一种新的深度学习框架TabNSA，它将原生稀疏注意（NSA）与TabMixer主干集成在一起，以有效地对表格数据建模。TabNSA通过动态关注每个实例的相关特征子集来解决计算和表示方面的挑战。NSA模块采用分层稀疏关注机制，包括令牌压缩、选择性保存和局部滑动窗口，在解决特征异质性的同时显著降低了标准关注操作的二次复杂度。作为补充，TabMixer主干通过具有独立参数的并行多层感知器（MLP）分支捕获复杂的非线性依赖关系。这些模块通过元素求和和均值池协同组合，使TabNSA能够建模全局上下文和细粒度交互。在监督学习和迁移学习设置中进行的大量实验表明，TabNSA始终优于最先进的深度学习模型。此外，通过使用微调的大型语言模型（LLM）来增强TabNSA，我们使其能够通过语言引导的对各种表格基准的泛化来有效地解决Few-Shot Learning挑战。代码可在：https://github.com/aseslamian/TabNSA。

{"title":"TabNSA: Native sparse attention for efficient tabular data learning","authors":"Ali Eslamian , Qiang Cheng","doi":"10.1016/j.neucom.2026.132928","DOIUrl":"10.1016/j.neucom.2026.132928","url":null,"abstract":"<div><div>Tabular data poses unique challenges for deep learning due to its heterogeneous feature types, lack of spatial structure, and often limited sample sizes. We propose TabNSA, a novel deep learning framework that integrates Native Sparse Attention (NSA) with a TabMixer backbone to efficiently model tabular data. TabNSA tackles computational and representational challenges by dynamically focusing on relevant feature subsets per instance. The NSA module employs a hierarchical sparse attention mechanism, including token compression, selective preservation, and localized sliding windows, to significantly reduce the quadratic complexity of standard attention operations while addressing feature heterogeneity. Complementing this, the TabMixer backbone captures complex, non-linear dependencies through parallel multilayer perceptron (MLP) branches with independent parameters. These modules are synergistically combined via element-wise summation and mean pooling, enabling TabNSA to model both global context and fine-grained interactions. Extensive experiments across supervised and transfer learning settings show that TabNSA consistently outperforms state-of-the-art deep learning models. Furthermore, by augmenting TabNSA with a fine-tuned large language model (LLM), we enable it to effectively address Few-Shot Learning challenges through language-guided generalization on diverse tabular benchmarks. <strong>Code available on:</strong> <span><span>https://github.com/aseslamian/TabNSA</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132928"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

WAVE++: Capturing within-task variance for continual relation extraction with adaptive prompting 捕获任务内方差，通过自适应提示持续提取关系

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132915

Bao-Ngoc Dao , Minh Le , Quang Nguyen , Luyen Ngo Dinh , Nam Le Hai, Linh Ngo Van

Memory-based approaches have shown strong performance in Continual Relation Extraction (CRE). However, storing examples from previous tasks increases memory usage and raises privacy concerns. Recently, prompt-based methods have emerged as a promising alternative, as they do not rely on storing past samples. Despite this progress, current prompt-based techniques face several core challenges in CRE, particularly in accurately identifying task identities and mitigating catastrophic forgetting. Existing prompt selection strategies often suffer from inaccuracies, lack robust mechanisms to prevent forgetting in shared parameters, and struggle to handle both cross-task and within-task variations. In this paper, we propose WAVE++, a novel approach inspired by the connection between prefix-tuning and mixture of experts. Specifically, we introduce task-specific prompt pools that enhance flexibility and adaptability across diverse tasks while avoiding boundary-spanning risks; this design more effectively captures both within-task and cross-task variations. To further refine relation classification, we incorporate label descriptions that provide richer, more global context, enabling the model to better distinguish among different relations. We also propose a training-free mechanism to improve task prediction during inference. Moreover, we integrate a generative model to consolidate prior knowledge within the shared parameters, thereby removing the need for explicit data storage. Extensive experiments demonstrate that WAVE++ outperforms state-of-the-art prompt-based and rehearsal-based methods, offering a more robust solution for continual relation extraction. Our code is publicly available at https://github.com/PiDinosauR2804/WAVE-CRE-PLUS-PLUS.

基于记忆的方法在连续关系提取（CRE）中表现出了很强的性能。但是，存储以前任务中的示例会增加内存使用并引起隐私问题。最近，基于提示的方法已经成为一种有前途的替代方法，因为它们不依赖于存储过去的样本。尽管取得了这些进展，但目前基于提示的技术在CRE中面临着几个核心挑战，特别是在准确识别任务身份和减轻灾难性遗忘方面。现有的提示选择策略往往存在不准确性，缺乏强大的机制来防止共享参数的遗忘，并且难以处理跨任务和任务内的变化。在本文中，我们提出了wav++，这是一种新颖的方法，灵感来自前缀调优和专家混合之间的联系。具体来说，我们引入了特定于任务的提示池，增强了跨不同任务的灵活性和适应性，同时避免了跨边界风险；这种设计更有效地捕获任务内和跨任务的变化。为了进一步改进关系分类，我们合并了标签描述，提供更丰富、更全局的上下文，使模型能够更好地区分不同的关系。我们还提出了一种无需训练的机制来改进推理过程中的任务预测。此外，我们集成了一个生成模型来整合共享参数中的先验知识，从而消除了显式数据存储的需要。大量的实验表明，wave++优于最先进的基于提示和基于预演的方法，为连续关系提取提供了更强大的解决方案。我们的代码可以在https://github.com/PiDinosauR2804/WAVE-CRE-PLUS-PLUS上公开获得。

{"title":"WAVE++: Capturing within-task variance for continual relation extraction with adaptive prompting","authors":"Bao-Ngoc Dao , Minh Le , Quang Nguyen , Luyen Ngo Dinh , Nam Le Hai, Linh Ngo Van","doi":"10.1016/j.neucom.2026.132915","DOIUrl":"10.1016/j.neucom.2026.132915","url":null,"abstract":"<div><div>Memory-based approaches have shown strong performance in Continual Relation Extraction (CRE). However, storing examples from previous tasks increases memory usage and raises privacy concerns. Recently, prompt-based methods have emerged as a promising alternative, as they do not rely on storing past samples. Despite this progress, current prompt-based techniques face several core challenges in CRE, particularly in accurately identifying task identities and mitigating catastrophic forgetting. Existing prompt selection strategies often suffer from inaccuracies, lack robust mechanisms to prevent forgetting in shared parameters, and struggle to handle both cross-task and within-task variations. In this paper, we propose <strong>WAVE++</strong>, a novel approach inspired by the connection between prefix-tuning and mixture of experts. Specifically, we introduce task-specific prompt pools that enhance flexibility and adaptability across diverse tasks while avoiding boundary-spanning risks; this design more effectively captures both within-task and cross-task variations. To further refine relation classification, we incorporate label descriptions that provide richer, more global context, enabling the model to better distinguish among different relations. We also propose a training-free mechanism to improve task prediction during inference. Moreover, we integrate a generative model to consolidate prior knowledge within the shared parameters, thereby removing the need for explicit data storage. Extensive experiments demonstrate that WAVE++ outperforms state-of-the-art prompt-based and rehearsal-based methods, offering a more robust solution for continual relation extraction. Our code is publicly available at <span><span>https://github.com/PiDinosauR2804/WAVE-CRE-PLUS-PLUS</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132915"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

The double-edged sword: A critical review of foundational medical datasets for AI benchmarks, biases, and the future of equitable healthcare 双刃剑：对人工智能基准、偏见和公平医疗的未来的基础医疗数据集进行批判性审查

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132919

Rabie A. Ramadan , Nadim K.M. Madi , Sallam O.F. Khairy , Kamal Aldin Yousif , Muataz Salam Al-Daweri , Alrajhi Waleed Khalid

The advancement of Artificial Intelligence (AI) has revolutionized medical diagnostics and treatment. Large-scale public datasets are fueling research in this field. Therefore, this systematic review is a comprehensive analysis of 13 foundational medical datasets. It evaluates the characteristics, performance metrics, and inherent biases of datasets across medical imaging, electronic health records, and genomics. The published literature is systematically reviewed to categorize these datasets, with a focus on performance metrics for everyday machine learning tasks. Additionally, this research documents evidence of systemic bias and limitations that affect model generalizability and clinical equity. Our analysis reveals compelling evidence that significant limitations temper the remarkable progress of algorithms. It has been frequently observed that AI models suffer dramatic accuracy drops when tested beyond their training distribution, with the Area Under the Curve consistently declining from 0.95 to 0.63. The research also identified consistent patterns of systemic bias that threaten the equitable application of healthcare. This bias stems from unrepresentative sampling, subjective annotation practices, label noise, and Natural Language Processing-derived ground-truth labels. Our findings demonstrate the urgent need for a paradigm shift in the development of medical applications. The AI and medical communities must prioritize generating diverse datasets and mitigating systematic bias. This study provides evidence-based recommendations and a technical toolkit to address these challenges and reduce any health disparities.

人工智能（AI）的进步彻底改变了医疗诊断和治疗。大规模的公共数据集正在推动这一领域的研究。因此，本系统综述是对13个基础医学数据集的综合分析。它评估了医学成像、电子健康记录和基因组学数据集的特征、性能指标和固有偏差。系统地回顾已发表的文献，对这些数据集进行分类，重点关注日常机器学习任务的性能指标。此外，本研究记录了影响模型普遍性和临床公平性的系统性偏见和局限性的证据。我们的分析揭示了令人信服的证据，表明显著的局限性抑制了算法的显著进步。人们经常观察到，当测试超出其训练分布时，人工智能模型的准确性会急剧下降，曲线下面积（Area Under the Curve）从0.95持续下降到0.63。该研究还确定了系统性偏见的一致模式，威胁到医疗保健的公平应用。这种偏差源于非代表性采样、主观注释实践、标签噪声和自然语言处理派生的基础真值标签。我们的研究结果表明，迫切需要在医学应用的发展模式转变。人工智能和医学界必须优先考虑生成多样化的数据集和减轻系统性偏见。这项研究提供了基于证据的建议和技术工具包，以应对这些挑战并减少任何健康差距。

{"title":"The double-edged sword: A critical review of foundational medical datasets for AI benchmarks, biases, and the future of equitable healthcare","authors":"Rabie A. Ramadan , Nadim K.M. Madi , Sallam O.F. Khairy , Kamal Aldin Yousif , Muataz Salam Al-Daweri , Alrajhi Waleed Khalid","doi":"10.1016/j.neucom.2026.132919","DOIUrl":"10.1016/j.neucom.2026.132919","url":null,"abstract":"<div><div>The advancement of Artificial Intelligence (AI) has revolutionized medical diagnostics and treatment. Large-scale public datasets are fueling research in this field. Therefore, this systematic review is a comprehensive analysis of 13 foundational medical datasets. It evaluates the characteristics, performance metrics, and inherent biases of datasets across medical imaging, electronic health records, and genomics. The published literature is systematically reviewed to categorize these datasets, with a focus on performance metrics for everyday machine learning tasks. Additionally, this research documents evidence of systemic bias and limitations that affect model generalizability and clinical equity. Our analysis reveals compelling evidence that significant limitations temper the remarkable progress of algorithms. It has been frequently observed that AI models suffer dramatic accuracy drops when tested beyond their training distribution, with the Area Under the Curve consistently declining from 0.95 to 0.63. The research also identified consistent patterns of systemic bias that threaten the equitable application of healthcare. This bias stems from unrepresentative sampling, subjective annotation practices, label noise, and Natural Language Processing-derived ground-truth labels. Our findings demonstrate the urgent need for a paradigm shift in the development of medical applications. The AI and medical communities must prioritize generating diverse datasets and mitigating systematic bias. This study provides evidence-based recommendations and a technical toolkit to address these challenges and reduce any health disparities.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132919"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

Memory recall-driven multi-view semantic inference for offensive language detection 基于记忆回忆的多视图语义推理攻击性语言检测

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132948

Zhiqiang Zhang , Tianpeng Cheng , Bing Li , Yuankang Sun , Chengxu Wang

The detection of offensive language plays a critical role in maintaining the health of online communities, preventing cyberbullying, and fostering inclusive communication. Current approaches utilize facilitated LLMs for direct aggressiveness classification, but flaws in complex contextual reasoning and in the detection of subtle cues in conversational environments greatly reduce detection performance. To address the aforementioned challenges, we propose the Memory Recall-Driven Multi-View Semantic Inference (MR-MVSI) model. Specifically, we first build a multi-view semantic inference module that enables the model to effectively capture subtle contextual cues and underlying emotional features from situational backgrounds, communicative targets, and emotions. Meanwhile, we employ a self-check mechanism to discriminate and regenerate the generated information, thereby ensuring the rigor and reliability of the inference process. In addition, we introduce a training memory recall module, which embeds the input samples into a highly semantic space and retrieves the most relevant memory segments to interpret complex linguistic patterns, thus significantly improving the detection accuracy. The experimental results demonstrate that our proposed MR-MVSI model achieves superior performance across all three benchmark datasets (OLID, HateXplain, and HatEval), with performance improvements of

6.6 %

,

0.2 %

, and

7.6 %

respectively.

攻击性语言的检测对于维护网络社区的健康、防止网络欺凌、促进包容性交流具有至关重要的作用。目前的方法利用便利的llm进行直接攻击性分类，但复杂的上下文推理和会话环境中微妙线索的检测缺陷大大降低了检测性能。为了解决上述问题，我们提出了记忆回忆驱动的多视图语义推理（MR-MVSI）模型。具体来说，我们首先构建了一个多视图语义推理模块，使模型能够有效地从情景背景、交际目标和情绪中捕获微妙的上下文线索和潜在的情感特征。同时，我们采用自检机制对生成的信息进行判别和再生，从而保证了推理过程的严谨性和可靠性。此外，我们引入了训练记忆召回模块，该模块将输入样本嵌入到高度语义空间中，并检索最相关的记忆片段来解释复杂的语言模式，从而显着提高了检测精度。实验结果表明，我们提出的MR-MVSI模型在所有三个基准数据集（OLID， HateXplain和HatEval）上都取得了优异的性能，性能分别提高了6.6%，0.2%和7.6%。

{"title":"Memory recall-driven multi-view semantic inference for offensive language detection","authors":"Zhiqiang Zhang , Tianpeng Cheng , Bing Li , Yuankang Sun , Chengxu Wang","doi":"10.1016/j.neucom.2026.132948","DOIUrl":"10.1016/j.neucom.2026.132948","url":null,"abstract":"<div><div>The detection of offensive language plays a critical role in maintaining the health of online communities, preventing cyberbullying, and fostering inclusive communication. Current approaches utilize facilitated LLMs for direct aggressiveness classification, but flaws in complex contextual reasoning and in the detection of subtle cues in conversational environments greatly reduce detection performance. To address the aforementioned challenges, we propose the <strong>M</strong>emory <strong>R</strong>ecall-Driven <strong>M</strong>ulti-<strong>V</strong>iew <strong>S</strong>emantic <strong>I</strong>nference (MR-MVSI) model. Specifically, we first build a multi-view semantic inference module that enables the model to effectively capture subtle contextual cues and underlying emotional features from situational backgrounds, communicative targets, and emotions. Meanwhile, we employ a self-check mechanism to discriminate and regenerate the generated information, thereby ensuring the rigor and reliability of the inference process. In addition, we introduce a training memory recall module, which embeds the input samples into a highly semantic space and retrieves the most relevant memory segments to interpret complex linguistic patterns, thus significantly improving the detection accuracy. The experimental results demonstrate that our proposed MR-MVSI model achieves superior performance across all three benchmark datasets (OLID, HateXplain, and HatEval), with performance improvements of <span><math><mn>6.6</mn><mi>%</mi></math></span>, <span><math><mn>0.2</mn><mi>%</mi></math></span>, and <span><math><mn>7.6</mn><mi>%</mi></math></span> respectively.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132948"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

MODE+: A benchmark and a probe into multimodal open-domain dialogue evaluation MODE+：对多模态开放域对话评估的基准和探索

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132787

Hang Yin , Xinglin Wang , Yueqi Zhang, Pinren Lu, Bin Sun, Peiwen Yuan, Kan Li

Multimodal Open-domain Dialogue (MOD) plays a crucial role in AI-human interactions and has garnered substantial interest. Although existing studies have explored various aspects of MOD, the evaluation of MOD remains underexplored. In this work, we propose MODE+, an evaluation benchmark for MOD and a probe into multimodel open-domain dialogue evaluation. Specifically, we construct MODE with a balanced difficulty distribution and divide it into three parts: MODE-Base and MODE-Hard, both consisting of single-turn dialogues, with MODE-Base containing 889 test cases and MODE-Hard comprising 215 more challenging cases designed for probing model robustness against multimodal inconsistencies. Additionally, we include MODE-Multi, which contains over 10,000 multi-turn dialogue cases for more extensive testing. Each case contains an image, a context, and turn-level response scores provided by at least three human annotators following standardized criteria. The consistency of human annotations has an average Spearman correlation of over 0.9, indicating that MODE is highly reliable in annotation. We test the MOD evaluation capabilities of various evaluators on MODE, including LLaMA, Claude3, GPT-4, LLaVA, Gemini and Qwen3-VL. Results show that even the best-performing model-based evaluators have surprisingly low agreement with human evaluations, with consistency scores for MODE-Base below 0.7 and for MODE-Hard falling below 0.4. To improve model-based MOD evaluation capabilities, we propose the MM-Eval framework, a systematic methodology designed to standardize automatic evaluation. MM-Eval introduces Image Transformation as a modality-bridging mechanism, Inference Enhancement for transparent reasoning, and Inference Calibration for statistical reliability. Compared to the baselines, MM-Eval achieves a 67.41% improvement on MODE-Base and a 297% enhancement on MODE-Hard. Furthermore, the performance on MODE-Multi shows significant improvements with MM-Eval, demonstrating that the framework is capable of handling larger and more complex datasets. These results demonstrate that MM-Eval serves as a transferable and robust standard for future MOD evaluation.

多模态开放域对话（multi - modal Open-domain Dialogue， MOD）在人工智能与人类交互中起着至关重要的作用，已经引起了人们的广泛关注。虽然已有的研究已经对MOD的各个方面进行了探索，但对MOD的评价仍存在不足。在这项工作中，我们提出了MOD的评估基准MODE+，并对多模型开放域对话评估进行了探索。具体来说，我们构建了具有平衡难度分布的MODE，并将其分为三部分：MODE- base和MODE- hard，均由单回合对话组成，MODE- base包含889个测试用例，MODE- hard包含215个更具挑战性的用例，旨在探索模型对多模态不一致性的鲁棒性。此外，我们还包括MODE-Multi，它包含超过10,000个多回合对话案例，用于更广泛的测试。每个案例包含一个图像、一个上下文和由至少三个人类注释者按照标准化标准提供的回合级响应分数。人工标注的一致性平均Spearman相关系数大于0.9，说明MODE在标注上具有较高的可靠性。我们测试了包括LLaMA、Claude3、GPT-4、LLaVA、Gemini和Qwen3-VL在内的多种评估器在MODE上的MOD评估能力。结果表明，即使是表现最好的基于模型的评估器与人类评估的一致性也低得惊人，MODE-Base的一致性得分低于0.7，MODE-Hard的一致性得分低于0.4。为了提高基于模型的MOD评估能力，我们提出了MM-Eval框架，这是一种系统化的方法，旨在规范自动评估。MM-Eval引入了图像转换作为模态桥接机制，用于透明推理的推理增强和用于统计可靠性的推理校准。与基线相比，MM-Eval在MODE-Base上提高了67.41%，在MODE-Hard上提高了297%。此外，MM-Eval在MODE-Multi上的性能得到了显著改善，表明该框架能够处理更大、更复杂的数据集。这些结果表明，MM-Eval可作为未来国防部评估的可转移和稳健的标准。

{"title":"MODE+: A benchmark and a probe into multimodal open-domain dialogue evaluation","authors":"Hang Yin , Xinglin Wang , Yueqi Zhang, Pinren Lu, Bin Sun, Peiwen Yuan, Kan Li","doi":"10.1016/j.neucom.2026.132787","DOIUrl":"10.1016/j.neucom.2026.132787","url":null,"abstract":"<div><div>Multimodal Open-domain Dialogue (MOD) plays a crucial role in AI-human interactions and has garnered substantial interest. Although existing studies have explored various aspects of MOD, the evaluation of MOD remains underexplored. In this work, we propose MODE+, an evaluation benchmark for MOD and a probe into multimodel open-domain dialogue evaluation. Specifically, we construct MODE with a balanced difficulty distribution and divide it into three parts: MODE-Base and MODE-Hard, both consisting of single-turn dialogues, with MODE-Base containing 889 test cases and MODE-Hard comprising 215 more challenging cases designed for probing model robustness against multimodal inconsistencies. Additionally, we include MODE-Multi, which contains over 10,000 multi-turn dialogue cases for more extensive testing. Each case contains an image, a context, and turn-level response scores provided by at least three human annotators following standardized criteria. The consistency of human annotations has an average Spearman correlation of over 0.9, indicating that MODE is highly reliable in annotation. We test the MOD evaluation capabilities of various evaluators on MODE, including LLaMA, Claude3, GPT-4, LLaVA, Gemini and Qwen3-VL. Results show that even the best-performing model-based evaluators have surprisingly low agreement with human evaluations, with consistency scores for MODE-Base below 0.7 and for MODE-Hard falling below 0.4. To improve model-based MOD evaluation capabilities, we propose the MM-Eval framework, a systematic methodology designed to standardize automatic evaluation. MM-Eval introduces Image Transformation as a modality-bridging mechanism, Inference Enhancement for transparent reasoning, and Inference Calibration for statistical reliability. Compared to the baselines, MM-Eval achieves a 67.41% improvement on MODE-Base and a 297% enhancement on MODE-Hard. Furthermore, the performance on MODE-Multi shows significant improvements with MM-Eval, demonstrating that the framework is capable of handling larger and more complex datasets. These results demonstrate that MM-Eval serves as a transferable and robust standard for future MOD evaluation.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132787"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147508","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

Hi-CBM: Mitigating information leakage via hierarchical concept bottleneck modeling Hi-CBM：通过分层概念瓶颈建模减少信息泄漏

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132916

Ao Sun , Qingyue Wang , Yuanyuan Yuan , Pingchuan Ma , Shuai Wang

Concept Bottleneck Models (CBMs) enhance interpretability and facilitate effective intervention by explicitly mapping input features to labels through human-understandable concepts. However, existing CBM frameworks often suffer from information leakage, wherein latent unintended information bypasses the concept layer, undermining interpretability and contaminating downstream predictions. To address this challenge, we propose Hi-CBM, a refined CBM framework that explicitly safeguards the two inherent mappings in CBMs—

features \to concepts

and

concepts \to classes

—to prevent leakage. Specifically, a Concept-Bottleneck Pooling mechanism regulates the feature-to-concept mapping by selectively aggregating latent features into semantic concepts to filter out irrelevant signals, while a binary Intervention Matrix governs the concept-to-class mapping by constraining concept–class associations, preventing unintended information encoded in concept probabilities from influencing final predictions. Extensive experiments across multiple datasets show that Hi-CBM substantially mitigates information leakage and produces concept representations that are both interpretable and intervenable, while maintaining strong predictive performance.

概念瓶颈模型（CBMs）通过人类可理解的概念显式地将输入特征映射到标签，从而增强了可解释性，并促进了有效的干预。然而，现有的CBM框架经常遭受信息泄漏，其中潜在的意外信息绕过概念层，破坏可解释性并污染下游预测。为了应对这一挑战，我们提出了Hi-CBM，这是一个改进的CBM框架，它明确地保护了CBM中的两个固有映射——特征→概念和概念→类——以防止泄漏。具体来说，概念-瓶颈池机制通过选择性地将潜在特征聚合到语义概念中以过滤掉无关信号来调节特征到概念的映射，而二元干预矩阵通过约束概念-类关联来控制概念到类的映射，防止概念概率中编码的意外信息影响最终预测。跨多个数据集的广泛实验表明，Hi-CBM大大减轻了信息泄漏，并产生了可解释和可干预的概念表示，同时保持了强大的预测性能。

{"title":"Hi-CBM: Mitigating information leakage via hierarchical concept bottleneck modeling","authors":"Ao Sun , Qingyue Wang , Yuanyuan Yuan , Pingchuan Ma , Shuai Wang","doi":"10.1016/j.neucom.2026.132916","DOIUrl":"10.1016/j.neucom.2026.132916","url":null,"abstract":"<div><div>Concept Bottleneck Models (CBMs) enhance interpretability and facilitate effective intervention by explicitly mapping input features to labels through human-understandable concepts. However, existing CBM frameworks often suffer from <em>information leakage</em>, wherein latent unintended information bypasses the concept layer, undermining interpretability and contaminating downstream predictions. To address this challenge, we propose Hi-CBM, a refined CBM framework that explicitly safeguards the two inherent mappings in CBMs—<span><math><mtext>features</mtext><mo>→</mo><mtext>concepts</mtext></math></span> and <span><math><mtext>concepts</mtext><mo>→</mo><mtext>classes</mtext></math></span>—to prevent leakage. Specifically, a <em>Concept-Bottleneck Pooling</em> mechanism regulates the feature-to-concept mapping by selectively aggregating latent features into semantic concepts to filter out irrelevant signals, while a binary <em>Intervention Matrix</em> governs the concept-to-class mapping by constraining concept–class associations, preventing unintended information encoded in concept probabilities from influencing final predictions. Extensive experiments across multiple datasets show that Hi-CBM substantially mitigates information leakage and produces concept representations that are both interpretable and intervenable, while maintaining strong predictive performance.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132916"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147234","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

TFIGF: Fire data augmentation model based on text-to-image synthesis TFIGF：基于文本图像合成的火灾数据增强模型

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-05 DOI: 10.1016/j.neucom.2026.132912

Hongyang Zhao , Yanan Guo , Xingdong Li , Yi Liu , Jing Jin

Data imbalance is one of the most challenging issues in deep learning, particularly in the domain of fire detection. In this field, the number of non-fire images significantly exceeds that of fire images, and the diversity of background information in images poses substantial challenges to fire detection. Recently, there have been significant advancements in generating images from textual descriptions using large language models. Inspired by this progress, this paper proposes an innovative Text-to-Image Fire Image Generation Framework (TFIGF). This framework aims to address the problem of insufficient model training due to a lack of adequate positive samples by generating fire images with varied backgrounds, thereby enhancing the efficiency and accuracy of fire detection. The proposed TFIGF framework consists of a front-end image generator and a back-end image filter. The image generator, comprising a feature fusion component, a CLIP image encoder based on the Vision Transformer (ViT), and a feature generation segment, is capable of merging textual information with the prior knowledge in the pre-trained CLIP-ViT model to produce images, enhancing the relevance and diversity of the generated images. Images produced by the image generator are evaluated and filtered by the image filter to obtain fire images most congruent with the textual descriptions. The proposed image filter converts the generated visual information into textual descriptions using ViT and GPT-3, and measures the alignment between the generated images and input text using cosine similarity. The proposed method can generate higher-quality images compared to state-of-the-art generative image methods. Furthermore, to verify the improvements in accuracy and reliability of fire detection with images generated by TFIGF, we constructed datasets augmented to various sizes, trained several popular detection models on these datasets, and tested them with real-world data. Experimental results demonstrate that images generated by TFIGF significantly enhance network performance in fire detection, confirming the framework’s potential and practicality in addressing data imbalance issues.

数据不平衡是深度学习中最具挑战性的问题之一，特别是在火灾探测领域。在该领域，非火灾图像的数量大大超过了火灾图像的数量，并且图像中背景信息的多样性给火灾检测带来了很大的挑战。最近，在使用大型语言模型从文本描述生成图像方面取得了重大进展。受此启发，本文提出了一种创新的文本到图像火焰图像生成框架（TFIGF）。该框架旨在通过生成不同背景的火灾图像，解决由于缺乏足够的正样本而导致的模型训练不足的问题，从而提高火灾探测的效率和准确性。提出的TFIGF框架由前端图像生成器和后端图像滤波器组成。该图像生成器包括特征融合组件、基于视觉变换（Vision Transformer, ViT）的CLIP图像编码器和特征生成片段，能够将文本信息与预训练的CLIP-ViT模型中的先验知识合并生成图像，增强生成图像的相关性和多样性。图像生成器生成的图像由图像过滤器进行评估和过滤，以获得与文本描述最一致的5张图像。该图像滤波器利用ViT和GPT-3将生成的视觉信息转换为文本描述，并利用余弦相似度度量生成的图像与输入文本的对齐程度。与最先进的生成图像方法相比，所提出的方法可以生成更高质量的图像。此外，为了验证使用TFIGF生成的图像进行火灾探测的准确性和可靠性的提高，我们构建了扩充到不同大小的数据集，在这些数据集上训练了几种流行的探测模型，并用实际数据对它们进行了测试。实验结果表明，TFIGF生成的图像显著提高了网络在火灾检测中的性能，证实了该框架在解决数据不平衡问题方面的潜力和实用性。

{"title":"TFIGF: Fire data augmentation model based on text-to-image synthesis","authors":"Hongyang Zhao , Yanan Guo , Xingdong Li , Yi Liu , Jing Jin","doi":"10.1016/j.neucom.2026.132912","DOIUrl":"10.1016/j.neucom.2026.132912","url":null,"abstract":"<div><div>Data imbalance is one of the most challenging issues in deep learning, particularly in the domain of fire detection. In this field, the number of non-fire images significantly exceeds that of fire images, and the diversity of background information in images poses substantial challenges to fire detection. Recently, there have been significant advancements in generating images from textual descriptions using large language models. Inspired by this progress, this paper proposes an innovative Text-to-Image Fire Image Generation Framework (TFIGF). This framework aims to address the problem of insufficient model training due to a lack of adequate positive samples by generating fire images with varied backgrounds, thereby enhancing the efficiency and accuracy of fire detection. The proposed TFIGF framework consists of a front-end image generator and a back-end image filter. The image generator, comprising a feature fusion component, a CLIP image encoder based on the Vision Transformer (ViT), and a feature generation segment, is capable of merging textual information with the prior knowledge in the pre-trained CLIP-ViT model to produce images, enhancing the relevance and diversity of the generated images. Images produced by the image generator are evaluated and filtered by the image filter to obtain fire images most congruent with the textual descriptions. The proposed image filter converts the generated visual information into textual descriptions using ViT and GPT-3, and measures the alignment between the generated images and input text using cosine similarity. The proposed method can generate higher-quality images compared to state-of-the-art generative image methods. Furthermore, to verify the improvements in accuracy and reliability of fire detection with images generated by TFIGF, we constructed datasets augmented to various sizes, trained several popular detection models on these datasets, and tested them with real-world data. Experimental results demonstrate that images generated by TFIGF significantly enhance network performance in fire detection, confirming the framework’s potential and practicality in addressing data imbalance issues.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132912"},"PeriodicalIF":6.5,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

Weightless multi-class classifier ruled by “Mental images” 由“心理意象”支配的失重多类分类器

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-04 DOI: 10.1016/j.neucom.2026.132975

Antonio Sorgente, Massimo De Gregorio

Common-sense rules represent a form of implicit knowledge derived from experience and observation, often used by both humans and artificial systems to guide decision-making. In this work, we introduce a novel extension of the DRASiW weightless neural network architecture that enables the automatic extraction of such rules from its internal knowledge representations, known as “Mental Images” (MIs). These rules, grounded in statistical patterns learned during training, are used to selectively inhibit a subset of discriminators prior to classification, effectively acting as a pre-filtering mechanism. The inhibition process is designed to reduce the computational burden and improve classification plausibility by narrowing the focus to the most promising classes. Unlike previous approaches that use MIs solely for interpretation or symbolic conversion, our method exploits this implicit knowledge to enhance the system’s own performance. The extracted rules reflect structural regularities within the input space and serve as a proxy for common sense reasoning embedded within the trained model.

This work highlights how internal representations in weightless neural systems can be explicitly leveraged to support interpretable and efficient decision processes, bridging the gap between subsymbolic learning and symbolic reasoning through rule-based inhibition.

常识性规则代表了一种从经验和观察中获得的隐性知识，通常被人类和人工系统用来指导决策。在这项工作中，我们引入了DRASiW无权重神经网络架构的新扩展，该架构能够从其内部知识表示（称为“心理图像”（MIs））中自动提取这些规则。这些规则基于训练期间学习到的统计模式，用于在分类之前选择性地抑制鉴别器子集，有效地充当预过滤机制。抑制过程旨在通过将焦点缩小到最有希望的类别来减少计算负担并提高分类的合理性。不像以前的方法只使用MIs进行解释或符号转换，我们的方法利用这种隐性知识来提高系统自身的性能。提取的规则反映了输入空间中的结构规律，并作为嵌入在训练模型中的常识推理的代理。这项工作强调了如何明确利用失重神经系统中的内部表征来支持可解释和有效的决策过程，通过基于规则的抑制弥合亚符号学习和符号推理之间的差距。

{"title":"Weightless multi-class classifier ruled by “Mental images”","authors":"Antonio Sorgente, Massimo De Gregorio","doi":"10.1016/j.neucom.2026.132975","DOIUrl":"10.1016/j.neucom.2026.132975","url":null,"abstract":"<div><div>Common-sense rules represent a form of implicit knowledge derived from experience and observation, often used by both humans and artificial systems to guide decision-making. In this work, we introduce a novel extension of the DRASiW weightless neural network architecture that enables the automatic extraction of such rules from its internal knowledge representations, known as “Mental Images” (MIs). These rules, grounded in statistical patterns learned during training, are used to selectively inhibit a subset of discriminators prior to classification, effectively acting as a pre-filtering mechanism. The inhibition process is designed to reduce the computational burden and improve classification plausibility by narrowing the focus to the most promising classes. Unlike previous approaches that use MIs solely for interpretation or symbolic conversion, our method exploits this implicit knowledge to enhance the system’s own performance. The extracted rules reflect structural regularities within the input space and serve as a proxy for common sense reasoning embedded within the trained model.</div><div>This work highlights how internal representations in weightless neural systems can be explicitly leveraged to support interpretable and efficient decision processes, bridging the gap between subsymbolic learning and symbolic reasoning through rule-based inhibition.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132975"},"PeriodicalIF":6.5,"publicationDate":"2026-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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

Edge-centric community hiding based on permanence in attributed networks 基于属性网络持久性的边缘中心社区隐藏

IF 6.5 2区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Neurocomputing

Pub Date : 2026-02-04 DOI: 10.1016/j.neucom.2026.132924

Zhichao Feng , Bohan Zhang , Junchang Jing , Dong Liu

Attributed networks contain both structural connections and rich node attributes, which are crucial for the formation and identification of community structures. Although integrating attribute data enhances the accuracy of community detection algorithms, it also raises the risk of privacy leakage. To address this issue, community hiding has emerged as a promising solution. However, most existing research has centered on topological networks, leaving attributed networks largely unexplored. In response to these issues, we propose Attribute Permanence (APERM)—a novel community hiding method specifically designed for attributed networks, which quantifies permanence loss to identify structurally influential edges for perturbation. The objective of our perturbation strategy is to disrupt the global community structure, which typically involves considering all existing and potential edges in the network, and this introduces considerable computational complexity. To tackle this problem, we introduce a strategy that identifies Closely Homogeneous Nodes (CHN) by integrating both structural similarity and attribute information, thereby significantly reducing the edge perturbation search space. The experimental results from eight community detection algorithms (four for attributed networks and four for non-attributed networks) across six real-world datasets demonstrate that our proposed APERM algorithm not only achieves effective community hiding but also retains robust performance.

属性网络既包含结构连接，又包含丰富的节点属性，这对社区结构的形成和识别至关重要。虽然整合属性数据提高了社区检测算法的准确性，但也增加了隐私泄露的风险。为了解决这个问题，社区隐藏已经成为一个有希望的解决方案。然而，大多数现有的研究都集中在拓扑网络上，使得属性网络在很大程度上未被探索。针对这些问题，我们提出了属性持久性（APERM）——一种专门为属性网络设计的新型社区隐藏方法，该方法量化持久性损失以识别扰动的结构影响边。我们的扰动策略的目标是破坏全球社区结构，这通常涉及考虑网络中所有现有和潜在的边，这引入了相当大的计算复杂性。为了解决这一问题，我们引入了一种通过整合结构相似性和属性信息来识别紧密同构节点（CHN）的策略，从而显著减少了边缘扰动搜索空间。基于6个真实数据集的8种社区检测算法（4种用于属性网络，4种用于非属性网络）的实验结果表明，我们提出的APERM算法不仅实现了有效的社区隐藏，而且保持了鲁棒性。

{"title":"Edge-centric community hiding based on permanence in attributed networks","authors":"Zhichao Feng , Bohan Zhang , Junchang Jing , Dong Liu","doi":"10.1016/j.neucom.2026.132924","DOIUrl":"10.1016/j.neucom.2026.132924","url":null,"abstract":"<div><div>Attributed networks contain both structural connections and rich node attributes, which are crucial for the formation and identification of community structures. Although integrating attribute data enhances the accuracy of community detection algorithms, it also raises the risk of privacy leakage. To address this issue, community hiding has emerged as a promising solution. However, most existing research has centered on topological networks, leaving attributed networks largely unexplored. In response to these issues, we propose Attribute Permanence (APERM)—a novel community hiding method specifically designed for attributed networks, which quantifies permanence loss to identify structurally influential edges for perturbation. The objective of our perturbation strategy is to disrupt the global community structure, which typically involves considering all existing and potential edges in the network, and this introduces considerable computational complexity. To tackle this problem, we introduce a strategy that identifies Closely Homogeneous Nodes (CHN) by integrating both structural similarity and attribute information, thereby significantly reducing the edge perturbation search space. The experimental results from eight community detection algorithms (four for attributed networks and four for non-attributed networks) across six real-world datasets demonstrate that our proposed APERM algorithm not only achieves effective community hiding but also retains robust performance.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"675 ","pages":"Article 132924"},"PeriodicalIF":6.5,"publicationDate":"2026-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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