NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130577
Dantong Qin , Yang Long , Xun Zhang , Zhibin Zhou , Yuting Jin , Pan Wang
{"title":"Towards stereoscopic vision: Attention-guided gaze estimation with EEG in 3D space","authors":"Dantong Qin , Yang Long , Xun Zhang , Zhibin Zhou , Yuting Jin , Pan Wang","doi":"10.1016/j.neucom.2025.130577","DOIUrl":"10.1016/j.neucom.2025.130577","url":null,"abstract":"<div><div>Since traditional gaze-tracking methods rely on line-of-sight estimation, spatial attention modeling from neural activity offers an alternative perspective to gaze estimation. This paper presents a proof-of-concept study on attention-guided gaze estimation with Electroencephalography (EEG), investigating whether brain signals can be leveraged to estimate attentional focus within a controlled 3D environment. We first conducted a preliminary survey to gather public opinions, revealing a generally positive attitude towards EEG-driven gaze tracking. Building on this insight, we collected an EEG dataset in VR, where participants engaged with stimuli presented at predefined spatial locations. We introduce a deep learning model that estimates the relative saliency of candidate positions, enabling gaze estimation through optimization within the learned representation. Our results demonstrate that attentional focus was successfully mapped in a 3D coordinate space from 5 participants, and low-frequency oscillations contributed more significantly to predictive performance. The model achieved robust accuracy in distinguishing gaze locations, highlighting the potential of EEG-based gaze estimation for attention tracking in 3D environments.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"648 ","pages":"Article 130577"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144270500","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}
{"title":"DSAIS-PINN: Dynamic seeds allocation importance sampling for physics-informed neural networks","authors":"Wentao Feng, Ming Hu, Rui Huang, Chenwei Tang, Shudong Huang, Jiancheng Lv","doi":"10.1016/j.neucom.2025.130578","DOIUrl":"10.1016/j.neucom.2025.130578","url":null,"abstract":"<div><div>Physics-Informed Neural Networks (PINNs) have gained widespread recognition as a capable tool for solving partial differential equations (PDEs), achieving notable success across various applications. Conventional PINNs, on the other hand, encounter significant challenges, including slow convergence rates and substantial computational costs. Recent advancements have incorporated importance sampling methods into the training process of PINNs, leading to considerable improvements in convergence speed and overall efficiency. Despite these advances, current importance sampling techniques typically calculate precise sampling probabilities for seed points and interpolate for non-seed points, which may lead to inaccuracies in high-loss regions. This study introduces the Dynamic Seed Allocation Importance Sampling (DSAIS) method, which dynamically adjusts the number of seed points in regions, i.e., Voronoi cells, with varying loss levels, prioritizing high-loss areas. This refinement improves interpolation accuracy for non-seed points. The proposed method’s effectiveness is confirmed through its application to three benchmark issues: the Schrödinger equation, the 1D Burgers equation, the Korteweg–de Vries equation, and the 2D heat equation. Performance is assessed using standard evaluation metrics, including maximum error, mean absolute error, and root mean square error. The outcomes demonstrate that DSAIS achieves state-of-the-art performance. DSAIS shows up to 63%, 30%, 30% smaller root mean square error in Schrodinger equation, Burgers equation and KDV equation, respectively.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"648 ","pages":"Article 130578"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144270483","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130571
Nian Wang , Zhigao Cui , Aihua Li , Yuanliang Xue , Rong Wang , Feiping Nie
{"title":"Multi-order graph based clustering via dynamical low rank tensor approximation","authors":"Nian Wang , Zhigao Cui , Aihua Li , Yuanliang Xue , Rong Wang , Feiping Nie","doi":"10.1016/j.neucom.2025.130571","DOIUrl":"10.1016/j.neucom.2025.130571","url":null,"abstract":"<div><div>Graph based clustering involves learning a proximity matrix with explicit clustering structure. However, since the limited link inputs and insufficient graph fusion, current works always obtain poor graphs with suboptimal clustering results. To solve the problem, in this paper, we propose a novel multi-order graph based clustering model via dynamic low-rank tensor approximation (MCDLT). Firstly, we use high-order proximity to enrich the link relations for graph inputs. Then a graph selection mechanism and low rank tensor approximation method are used to dynamically learn the consistent information from the complex links in various-order graphs. Finally, we propose a doubly stochastic graph fusion method to directly learn a symmetrical graph that provides clustering results by its connectivity. A novel Augmented Lagrangian Multiplier (ALM) based method is proposed for the sub-problem of doubly stochastic constraints. Experiments shows that our method learns a graph with clearer data structure, achieving SOTA clustering performance and obtaining the GroundTruth for JAFFE data set. Code will be published at <span><span>https://github.com/NianWang-HJJGCDX/MCDLT.git</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130571"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144254769","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130568
Luyao Chang, Leiting Chen, Chuan Zhou
{"title":"Uncertainty-Aware Contrastive Learning for deep clustering","authors":"Luyao Chang, Leiting Chen, Chuan Zhou","doi":"10.1016/j.neucom.2025.130568","DOIUrl":"10.1016/j.neucom.2025.130568","url":null,"abstract":"<div><div>Deep clustering aims to group unlabeled data into meaningful clusters by learning discriminative feature representations. However, ambiguous features often lead to noisy representations and inconsistent semantics, limiting improvements in clustering performance. To address this issue, we propose an Uncertainty-Aware Contrastive Learning (UACL) method for deep clustering, which achieves robustness by adaptively restricting the learning of ambiguous features. Specifically, we model pairwise similarity evidence via subjective logic theory, formulating co-cluster probabilities as a Dirichlet distribution to quantify epistemic uncertainty from feature ambiguity. Guided by this uncertainty, we design a dynamic weight-updating strategy that progressively extracts information from potential positives, enhancing the model’s ability to learn discriminative representations and semantically consistent clusters. Furthermore, to enforce attribute consistency, we develop an Attribute Distribution Alignment module that aligns similarity and uncertainty. Extensive experiments on five benchmark datasets demonstrate UACL outperforms current state-of-the-art methods, with an improved ACC of 3.5% for CIFAR-100 and 7.0% for ImageNet-Dogs. The source code is available at: <span><span>https://github.com/YL616/UACL</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130568"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144254773","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130565
Giacomo Cappiello , Filippo Caruso
{"title":"Quantum AI for Alzheimer’s disease early screening","authors":"Giacomo Cappiello , Filippo Caruso","doi":"10.1016/j.neucom.2025.130565","DOIUrl":"10.1016/j.neucom.2025.130565","url":null,"abstract":"<div><div>Quantum machine learning is a new research field combining quantum information science and machine learning. Quantum computing technologies appear to be particularly well-suited for addressing problems in the health sector efficiently. They have the potential to handle large datasets more effectively than classical models and offer greater transparency and interpretability for clinicians. Alzheimer’s disease is a neurodegenerative brain disorder that mostly affects elderly people, causing important cognitive impairments. It is the most common cause of dementia and it has an effect on memory, thought, learning abilities and movement control. This type of disease has no cure, consequently an early diagnosis is fundamental for reducing its impact. The analysis of handwriting can be effective for diagnosing, as many researches have conjectured. The DARWIN (Diagnosis AlzheimeR WIth haNdwriting) dataset contains handwriting samples from people affected by Alzheimer’s disease and a group of healthy people. Here we apply quantum AI to this use-case. In particular, we use this dataset to test classical methods for classification and compare their performances with the ones obtained via quantum machine learning methods. We find that quantum methods generally perform better than classical methods.</div><div>Our results pave the way for future new quantum machine learning applications in early-screening diagnostics in the healthcare domain.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130565"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144261945","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130708
Zhenglin Zhang , Tengfei Wang , Zian Hu , Li-Zhuang Yang , Hai Li
{"title":"Multivariate time series approach integrating cross-temporal and cross-channel attention for dysarthria detection from speech","authors":"Zhenglin Zhang , Tengfei Wang , Zian Hu , Li-Zhuang Yang , Hai Li","doi":"10.1016/j.neucom.2025.130708","DOIUrl":"10.1016/j.neucom.2025.130708","url":null,"abstract":"<div><div>Speech analysis offers a non-invasive, low-cost approach to dysarthria detection. Studies have shown that the temporal correlations within speech signals and the interactions among the multidimensional feature variables derived from them can facilitate dysarthria detection. However, current studies either rely on pre-designed feature sets, which depend heavily on cumbersome feature engineering, or focus solely on spectral or high-dimensional audio vectors that capture temporal dependencies while neglecting the interactions between internal multivariate features. We propose an end-to-end method that utilizes audio pre-trained models as multivariate time series feature extractors, combined with InceptionTime and cross-temporal and cross-channel attention mechanisms, to fully capture temporal dependencies and interactions among variables within speech for accurate dysarthria detection. Results show that the proposed method achieves a detection accuracy of 92.06 % on a local Mandarin dysarthria dataset, which is at least 2.17 percentage points higher than previous studies, with the highest stability and the lowest time cost. Furthermore, it achieves an accuracy of 87.73 % on an external English dataset, demonstrating good cross-linguistic adaptability and generalizability. Additionally, experiments show that in connected speech tasks, structured tasks outperform unstructured ones in leveraging interactions, leading to more effective dysarthria detection. These findings validate the effectiveness of the proposed end-to-end dysarthria detection method, further advancing the development of speech analysis as a promising tool for dysarthria screening.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130708"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144242014","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130580
Xiaoli Zhu, Luyang Yu, Yurong Liu
{"title":"Sampled-data-based synchronization of matrix-weighted multi-layer complex networks via pinning control","authors":"Xiaoli Zhu, Luyang Yu, Yurong Liu","doi":"10.1016/j.neucom.2025.130580","DOIUrl":"10.1016/j.neucom.2025.130580","url":null,"abstract":"<div><div>This paper examines the synchronization issue for a class of matrix-weighted multi-layer complex networks. Both intra-layer and inter-layer synchronization are investigated by means of sampled-data-based pinning control. In order to better describe the multidimensional attributes of state connection, matrix coupling rather than the conventional scalar weights is introduced to specify the node connections within and between layers. To reduce the waste of communication resources and save the control cost, a sampled-data-based pinning control strategy is designed to reach the intra-layer and inter-layer synchronization of matrix-weighted multi-layer complex networks. To effectively implement the pinning control, it is assumed that for each unpinned node (layer), there exists at least one pinned node (layer) which has a path to the unpinned node (layer). In this context, some sufficient conditions are derived through the utilization of algebraic graph theory, the Lyapunov stability theory, and a modified Halanay inequality. Finally, one simulation example is provided to verify the effectiveness of the obtained theoretical results.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130580"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144242253","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}
{"title":"From continuous pre-training to alignment: A comprehensive toolkit for large language models in federated learning","authors":"Zhuo Zhang , Yukun Zhang , Guanzhong Chen , Lizhen Qu , Xun Zhou , Hui Wang , Zenglin Xu","doi":"10.1016/j.neucom.2025.130572","DOIUrl":"10.1016/j.neucom.2025.130572","url":null,"abstract":"<div><div>The rapid success of Large Language Models (LLMs) has unlocked vast potential for AI applications in privacy-sensitive domains. However, the traditional centralized training of LLMs poses significant challenges due to privacy concerns regarding collecting sensitive data from diverse sources. This paper offers a promising and privacy-enhancing solution for LLMs: collaboratively training LLMs via Federated Learning (FL) across multiple clients, eliminating the need for raw data transmission. To this end, we present F4LLM, a new and comprehensive toolbox that supports the entire Federated LLM pipeline, from Continuous pre-training to alignment and LLM evaluation. F4LLM employs gRPC as the communication protocol to support various widely-used FL algorithms, ensuring efficient development and benchmarking in geo-distributed FL environments. Moreover, F4LLM offers both open-form and closed-form evaluation options via the efficient inference tool vLLM. The source code and documentation are at <span><span>here</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130572"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144261948","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130726
Lili Yu , Xubing Zhang , Kai Wang
{"title":"Few-shot learning framework based on classifier and domain adaptive alignment for hyperspectral classification","authors":"Lili Yu , Xubing Zhang , Kai Wang","doi":"10.1016/j.neucom.2025.130726","DOIUrl":"10.1016/j.neucom.2025.130726","url":null,"abstract":"<div><div>The focus of cross-domain hyperspectral image (HSI) classification is on the model training phase, where the model can utilize a large number of labeled samples from the source domain (SD) to learn prior knowledge. And transfer the knowledge to the target domain (TD) classification task that contains a few labeled samples. Nevertheless, it is difficult to align the features of the SD and TD into a unified space using only a few samples to obtain domain-invariant feature representations. To address this problem, a cross-domain few-shot learning (FSL) framework based on classifier and domain-level adaptive alignment (CDLA) is proposed for HSI classification. The separation-selection-refinement strategy (SSRS) is used to design the parameter-sharing network, which can utilize features with a large amount of information and extract more representative fine-grained features. During the training stage, labels are determined using the covariance metric, and additional classifiers are added to design label consistency constraints to optimize the class prototype and strengthen the intra-class feature consistency. In addition, the domain adversarial network is used to mitigate the effects of the domain shift and learn collaborative representation. Comparative experiments are conducted on three public HSI datasets to validate the effectiveness of the CDLA model and its superiority in existing FSL methods for HSI classification.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130726"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144242248","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}
NeurocomputingPub Date : 2025-06-06DOI: 10.1016/j.neucom.2025.130576
Hongji Zhuang , Shufan Wu , Vladimir Yu. Razoumny , Yury N. Razoumny
{"title":"H∞ control for cooperative multi-agent systems: Event-triggered off-policy reinforcement learning approach","authors":"Hongji Zhuang , Shufan Wu , Vladimir Yu. Razoumny , Yury N. Razoumny","doi":"10.1016/j.neucom.2025.130576","DOIUrl":"10.1016/j.neucom.2025.130576","url":null,"abstract":"<div><div>This paper investigates the <span><math><msub><mrow><mi>H</mi></mrow><mrow><mi>∞</mi></mrow></msub></math></span> synchronization control problem for cooperative multi-agent systems using an event-triggered off-policy reinforcement learning (OffRL) approach. Based on the connection graph from a global optimization perspective, a system model is first constructed to reformulate the problem as solving the Hamilton–Jacobi–Isaacs (HJI) equation. To address the computation and communication challenges associated with extensive information exchanges in multi-agent systems, an event-triggered scheme is introduced. A triggering condition is proposed, and its feasibility is rigorously analyzed in terms of system stability and the exclusion of Zeno behavior. To solve the HJI equation, a model-free OffRL algorithm is derived from the Bellman equation, leveraging system dataset collection to bypass potential inaccuracies in the system dynamic model. Finally, the feasibility and effectiveness of the proposed algorithm are rigorously demonstrated through theoretical analysis and a simulation example.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"647 ","pages":"Article 130576"},"PeriodicalIF":5.5,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144241997","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}