Neural Networks最新文献

{"title":"Adaptive performance control of switched nonlinear systems under false data injection attacks and input saturation constraints","authors":"Wenjing He ,&nbsp;Youqing Wang ,&nbsp;Yukun Shi","doi":"10.1016/j.neunet.2026.108602","DOIUrl":"10.1016/j.neunet.2026.108602","url":null,"abstract":"<div><div>Considering the controller-actuator channel subjected to false data injection (FDI) attacks, this study proposes an adaptive performance control strategy for switched nonlinear systems under composite disturbances and input saturation. The strategy achieves disturbance rejection and a more flexible prescribed performance control. In addition, implicit disturbances that are coupled with the system states in switched nonlinear systems lie beyond the capability of existing disturbance observers. Moreover, conventional prescribed performance control methods become ineffective in the presence of input saturation. To overcome these challenges, a novel switched nonlinear disturbance observer is developed to reconstruct implicit mismatched disturbances. This observer employs dynamic high gain technique to address the unknown nonlinearity of disturbances. Additionally, an adaptive update law is designed to compensate for matched disturbances. Then, a modified fixed-time performance function is introduced to balance input saturation with tracking performance. To mitigate the adverse effects of FDI attacks, the neural network approximation technique is applied within the backstepping control process. Finally, the simulation results reveal that the proposed adaptive performance control strategy achieves effective tracking control of switched nonlinear systems.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108602"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145999372","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}

{"title":"Point-Deeponet: Predicting nonlinear fields on non-Parametric geometries under variable load conditions","authors":"Jangseop Park ,&nbsp;Namwoo Kang","doi":"10.1016/j.neunet.2026.108560","DOIUrl":"10.1016/j.neunet.2026.108560","url":null,"abstract":"<div><div>Nonlinear structural analyses in engineering often require extensive finite element simulations, limiting their applicability in design optimization and real-time control. Conventional deep learning surrogates often struggle with complex, non-parametric three-dimensional (3D) geometries and directionally varying loads. This work presents Point-DeepONet, an operator-learning-based surrogate that integrates PointNet into the DeepONet framework to learn a mapping from non-parametric geometries and variable load conditions to physical response fields. By leveraging PointNet to learn a geometric representation from raw point clouds, our model circumvents the need for manual parameterization. This geometric embedding is then synergistically fused with load conditions within the DeepONet architecture to accurately predict three-dimensional displacement and von Mises stress fields. Trained on a large-scale dataset, Point-DeepONet demonstrates high fidelity, achieving a coefficient of determination (R²) reaching 0.987 for displacement and 0.923 for von Mises stress. Furthermore, to rigorously validate its generalization capabilities, we conducted additional experiments on unseen, randomly oriented load directions, where the model maintained exceptional accuracy. Compared to nonlinear finite element analyses that require about 19.32 minutes per case, Point-DeepONet provides predictions in mere seconds-approximately 400 times faster-while maintaining excellent scalability. These findings, validated through extensive experiments and ablation studies, highlight the potential of Point-DeepONet to enable rapid, high-fidelity structural analyses for complex engineering workflows.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108560"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981539","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}

{"title":"AdaAlign: A unified solution for traditional and modern zero-shot sketch-based image retrieval","authors":"Mingrui Zhu,&nbsp;Fangzhou Wang,&nbsp;Xin Wei,&nbsp;Nannan Wang,&nbsp;Xinbo Gao","doi":"10.1016/j.neunet.2026.108586","DOIUrl":"10.1016/j.neunet.2026.108586","url":null,"abstract":"<div><div>Zero-shot sketch-based image retrieval (ZS-SBIR) is challenging due to the cross-domain nature of sketches and photos, as well as the semantic gap between seen and unseen classes. With the rapid advancements in modern large vision-language models (VLMs), traditional approaches that relied solely on small vision encoders have gradually been supplanted. However, both the traditional vision encoder-based methods and the modern VLMs-based methods have their limitations, and no unified approach effectively addresses both. In this paper, we present an effective <em>“Adaptation and Alignment (AdaAlign)”</em> approach to address the key challenges. Specifically, we insert lightweight Adapter or LoRA to learn new abstract concepts of the sketches and improve cross-domain representation capabilities, which helps alleviate domain heterogeneity. Then, we propose to directly align the learned image embedding with the more semantically rich text embedding within a distillation framework to bridge the semantic gap. This enables the model to learn more generalizable visual representations from linguistic semantic cues. We integrate our key innovations into both traditional small models (<em>e.g.</em>, ResNet50 or DINO-S) and modern VLMs (<em>e.g.</em>, SigLIP), resulting in state-of-the-art performance. Extensive experiments on three benchmark datasets demonstrate the superiority of our method in terms of retrieval accuracy and flexibility.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108586"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981632","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}

{"title":"DMDNet: Dual-branch multi-modal deep fusion network for V-D-T salient object detection","authors":"Yaoqi Sun ,&nbsp;Bin Wan ,&nbsp;Haibing Yin ,&nbsp;Yahong Chen","doi":"10.1016/j.neunet.2026.108579","DOIUrl":"10.1016/j.neunet.2026.108579","url":null,"abstract":"<div><div>In the multi-modal salient object detection task, depth or thermal features are often directly fused with visible feature during the encoding stage, which directly results in the fused encoder features containing a large amount of noise information and reducing the accuracy of detection. To address this challenge, in this paper, we propose a novel dual-branch multi-modal deep fusion network (DMDNet) where visible image serves as one branch, and depth and thermal images serve as another branch to achieve multi-modal feature fusion in the decoder phase. In the encoder phase, we apply two types of backbone networks to three modalities to ensure sufficient information extraction and design the modal interaction (MI) module to dig the complementarity between depth and thermal features. In the decoder phase, we propose the multi-scale feature perception (MFP) module and region optimization (RO) module in succession to mine and optimize the saliency region. After that, we introduce the dual-branch fusion (DF) module to integrate multi-modal feature in the bottom-to-top manner for generating final saliency map. DMDNet achieves superior performance on the VDT-2048 dataset, as verified by comprehensive experimental results.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108579"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981634","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}

{"title":"CoCoFR: Collaborative codebooks learning with soft matching strategy for blind face restoration","authors":"Teng Feng ,&nbsp;Junwei Xu ,&nbsp;Tao Huang ,&nbsp;Zhenyu Wang ,&nbsp;Fangfang Wu ,&nbsp;Weisheng Dong ,&nbsp;Xin Li ,&nbsp;Guangming Shi","doi":"10.1016/j.neunet.2026.108607","DOIUrl":"10.1016/j.neunet.2026.108607","url":null,"abstract":"<div><div>Blind Face Restoration (BFR) has garnered considerable attention for its practical applicability to recover high-quality (HQ) facial images from their degraded versions. Existing BFR methods primarily incorporate diverse priors to mitigate its ill-posed nature. Notably, the codebook prior, which aggregates facial representations from HQ images has achieved impressive results. However, two performance constraints remain: i) The reliance on a single spatial-domain codebook neglects the potential information in the frequency domain. ii) The commonly used feature-matching strategies overlook the valid information encapsulated within the low-quality (LQ) identity features. To address these issues, we propose CoCoFR, which learns collaborative codebooks in both spatial and frequency domains and implements adaptive matching between LQ and HQ features with a designed Dual Codebooks Cross Attention (DCCA) module. Additionally, benefiting from its global receptive fields and linear complexity, CoCoFR facilitates coarse-to-fine feature fusion via a simple yet effective state space model (Mamba)-based fusion block MFB. Extensive experiments on both synthetic and real-world datasets validate the superiority of our CoCoFR in terms of realness and fidelity compared to state-of-the-art methods.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108607"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039400","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}

{"title":"EEG-to-gait decoding via phase-aware representation learning","authors":"Xi Fu ,&nbsp;Weibang Jiang ,&nbsp;Rui Liu ,&nbsp;Gernot R. Müller-Putz ,&nbsp;Cuntai Guan","doi":"10.1016/j.neunet.2026.108608","DOIUrl":"10.1016/j.neunet.2026.108608","url":null,"abstract":"<div><div>Accurate decoding of lower-limb motion from EEG signals is essential for advancing brain-computer interface (BCI) applications in movement intent recognition and control. This study presents <strong>NeuroDyGait</strong>, a two-stage, phase-aware EEG-to-gait decoding framework that explicitly models temporal continuity and domain relationships. To address challenges of causal, phase-consistent prediction and cross-subject variability, Stage I learns semantically aligned EEG-motion embeddings via relative contrastive learning with a cross-attention-based metric, while Stage II performs domain relation-aware decoding through dynamic fusion of session-specific heads. Comprehensive experiments on two benchmark datasets (GED and FMD) show substantial gains over baselines, including a recent 2025 model EEG2GAIT. The framework generalizes to unseen subjects and maintains inference latency below 5 ms per window, satisfying real-time BCI requirements. Visualization of learned attention and phase-specific cortical saliency maps further reveals interpretable neural correlates of gait phases. Future extensions will target rehabilitation populations and multimodal integration.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108608"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146031541","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}

{"title":"MT-IDS: A multi-task information decoupling strategy for identifying lymph node metastasis in the mediastinal region","authors":"Wei Zhou ,&nbsp;Yining Xie ,&nbsp;Fengjiao Wang ,&nbsp;Jing Zhao ,&nbsp;Jiayi Ma","doi":"10.1016/j.neunet.2026.108541","DOIUrl":"10.1016/j.neunet.2026.108541","url":null,"abstract":"<div><div>Accurately identifying the mediastinal regions where metastatic lymph nodes are located is critical for the staging diagnosis of lung cancer. This identification task involves two distinct detection dimensions: mediastinal region identification and lymph node metastasis assessment. Traditional single-task image classification algorithms struggle to manage the interference between different classification dimensions within a single task. Existing multi-task learning methods struggle to balance the relationship between shared and task-specific features, and often fail to effectively fit the underlying data distributions and task characteristics during gradient adjustment. To address these challenges, we propose a Multi-Task Information Decoupling Strategy (MT-IDS). MT-IDS decomposes the main task into multiple auxiliary tasks along different feature dimensions, forming a unified multi-task system to optimize detection performance across tasks. A Dual-control Branch Routing Gate Mechanism (DBR) is employed in MT-IDS to compute the weighting of shared and task-specific features, thereby enabling more precise expert selection and feature extraction for each task. Additionally, a Dual-Dimensional Gradient Balancing Algorithm (DD-GB) is introduced in MT-IDS, whereby gradient balance is achieved through alignment of gradient directions and dynamic scaling of magnitudes, while the distribution of inter-task gradient characteristics is maintained. The significant advantages demonstrated by MT-IDS in both ablation and comparative experiments indicate its potential as an innovative solution for multi-dimensional medical image classification problems.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108541"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981715","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}

{"title":"Exploring cognitive workload recognition using CogRepLKNet with EEG-fMRI","authors":"Yang Shao ,&nbsp;Yueying Zhou ,&nbsp;Xuyun Wen ,&nbsp;Peiliang Gong ,&nbsp;Qun Dai ,&nbsp;Daoqiang Zhang","doi":"10.1016/j.neunet.2026.108575","DOIUrl":"10.1016/j.neunet.2026.108575","url":null,"abstract":"<div><div>Accurate multimodal Cognitive Workload Recognition (CWR) remains challenging due to the difficulty of modeling cross-modal relationships between Electroencephalography (EEG) and Functional Magnetic Resonance Imaging (fMRI) data. Additionally, the inherent heterogeneity of these physiological signals-each capturing distinct neural characteristics-complicates unified feature extraction. To address this, we propose CogRepLKNet, a universal re-parameterizable large-kernel convolutional neural network (CNN) designed for multimodal EEG-fMRI modeling. CogRepLKNet employs two parallel universal perception branches consisting of stacked large- and small-kernel CNNs and an adaptive gated attention fusion mechanism to jointly capture complementary temporal-spatial dynamics from both modalities. This design enables efficient feature integration with reduced computational complexity and fewer training samples compared to transformer-based approaches. Through input projections, the perception module enables universal feature extraction across various physiological signals without altering its architecture. Experiments on a self-constructed EEG-fMRI dataset demonstrate that CogRepLKNet achieves state-of-the-art performance, while ensuring low training complexity and easy portability. CogRepLKNet holds great potential for advancing multimodal applications in CWR. Our code is available at <span><span>https://github.com/prestyan/CogRepLKNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108575"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981717","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}

{"title":"Graph-Agnostic Linear Transformers","authors":"Zhiyu Guo ,&nbsp;Yang Liu ,&nbsp;Xiang Ao ,&nbsp;Yateng Tang ,&nbsp;Xinhuan Chen ,&nbsp;Xuehao Zheng ,&nbsp;Qing He","doi":"10.1016/j.neunet.2026.108595","DOIUrl":"10.1016/j.neunet.2026.108595","url":null,"abstract":"<div><div>Graph Transformers (GTs), as emerging foundational encoders for graph-structured data, have shown promising performance due to the integration of local graph structures with global attention mechanisms. However, the complex attention functions and their coupling with graph structures incur significant computational overhead, particularly in large-scale graphs. In this paper, we decouple graph structures from Transformers and propose the Graph-Agnostic Linear Transformer (GALiT). In GALiT, graph structures are solely utilized to denoise raw node features before training, as our findings reveal that these denoised features have integrated the main information of the graph structure and can replace it to guide Transformers. By excluding graph structures from the training and inference stages, GALiT serves as a graph-agnostic model which significantly reduces computational complexity. Additionally, we simplify the linear attention functions inherited from traditional Transformers, which further reduces computational overhead while still capturing the relationships between nodes. Through weighted combination, we integrate the denoised features into the attention mechanism, as our theoretical analysis reveals the key role of the synergy between linear attention and denoised features in enhancing representation diversity. Despite decoupling graph structures and simplifying attention mechanisms, our model surprisingly outperforms most GNNs and GTs on benchmark graphs. Experimental results indicate that GALiT achieves high efficiency while maintaining or even enhancing performance.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108595"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039238","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}

{"title":"ReTri: Progressive domain bridging via representation disentanglement and triple-level consistency-driven feature alignment for unsupervised domain adaptive medical image segmentation","authors":"Xiaoru Gao ,&nbsp;Guoyan Zheng","doi":"10.1016/j.neunet.2026.108564","DOIUrl":"10.1016/j.neunet.2026.108564","url":null,"abstract":"<div><div>Unsupervised domain adaptation (UDA) in medical image segmentation presents significant challenges due to substantial cross-domain disparities and the inherent absence of target domain annotations. In this study, to address these challenges, we propose an end-to-end progressive domain bridging framework based on representation disentanglement and triple-level consistency-driven feature alignment, referred to as ReTri, that synergistically integrates a representation disentanglement-based image alignment (RDIA) module with a novel triple-level consistency-driven feature alignment (TCFA) module. In particular, the RDIA module aims to establish an initial domain bridge by decoupling and aligning fundamental visual disparities through disentangled representation learning, while the novel TCFA module hierarchically bridges remaining cross-domain semantic discrepancies and feature distribution disparities via two novel consistency-driven alignment mechanisms: 1) attention-guided semantics-level consistency alignment, where we purposely design a bi-attentive semantic feature extraction (BSFE) component coupled with an attention-adaptive semantic consistency (ASC) loss function, facilitating dynamic alignment of high-level semantic representations across domains, and 2) multi-view dual-level mixing consistency alignment, consisting of Feature-Cut consistent self-ensembling (FCCS) and Trans-Cut consistent self-ensembling (TCCS) components. These two components operate within intermediate mixing spaces to ensure robust knowledge transfer through complementary feature- and prediction-level consistency regularization. Extensive experimental evaluations are conducted on four challenging datasets (Lumbar Spine CT-MR, Cardiac CT-MR, Cross-domain Echocardiography, and Multi-center Prostate MR) across seven UDA-based segmentation scenarios and two external validation scenarios. Our framework achieves superior performance over the best state-of-the-art (SOTA) methods on following UDA-based segmentation scenarios: +2.9% DSC for spine CT → MR segmentation, +3.6% and +2.4% DSC for bidirectional cardiac CT↔MR segmentation, +1.7% and +2.3% DSC for bidirectional cross-center cross-vendor Echocardiography (CAMUS↔EchoNet-Dynamic) segmentation, and +12.2% and +12.0% DSC for bidirectional multi-center prostate MR segmentation. The source code and the datasets are publicly available at <span><span>https://github.com/xiaorugao999/ReTri</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"198 ","pages":"Article 108564"},"PeriodicalIF":6.3,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981537","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}