Neurocomputing最新文献

{"title":"Locality-constrained robust discriminant non-negative matrix factorization for depression detection: An fNIRS study","authors":"Yushan Wu ,&nbsp;Jitao Zhong ,&nbsp;Lu Zhang ,&nbsp;Hele Liu ,&nbsp;Shuai Shao ,&nbsp;Bin Hu ,&nbsp;Hong Peng","doi":"10.1016/j.neucom.2024.128887","DOIUrl":"10.1016/j.neucom.2024.128887","url":null,"abstract":"<div><div>Major depressive disorder (MDD) is having an increasingly severe impact worldwide, which creates a pressing need for an efficient and objective method of depression detection. Functional near-infrared spectroscopy (fNIRS), which directly monitors changes in cerebral oxygenation, has become an important tool in depression research. Currently, feature extraction methods based on multi-channel fNIRS data often overlook the local structure of the data and the subsequent classification cost. To address these challenges, we introduce an innovative feature extraction algorithm, namely locality-constrained robust discriminant non-negative matrix factorization (LRDNMF). The algorithm incorporates <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mn>2</mn><mo>,</mo><mn>1</mn></mrow></msub></math></span> regularization, local coordinate constraints, within-class scatter distance, and total scatter distance, achieving a fusion of robustness, locality, and discrimination. LRDNMF enhances feature representation, reduces noise impact, and significantly boosts classification ability. Based on experimental results from 56 participants, LRDNMF achieves an accuracy of 90.55%, a recall of 91.48%, a precision of 90.46%, and an F1 score of 0.91 under full stimuli. These results outperform existing algorithms, validating the effectiveness of LRDNMF and demonstrating its significant potential in auxiliary diagnosis of depression.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"617 ","pages":"Article 128887"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142745309","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":"Extending the learning using privileged information paradigm to logistic regression","authors":"Mario Martínez-García ,&nbsp;Susana García-Gutierrez ,&nbsp;Lasai Barreñada ,&nbsp;Iñaki Inza ,&nbsp;Jose A. Lozano","doi":"10.1016/j.neucom.2024.128869","DOIUrl":"10.1016/j.neucom.2024.128869","url":null,"abstract":"<div><div>Learning using privileged information paradigm is a learning scenario that exploits privileged features, available at training time, but not at prediction, as additional information for training models. This paper delves into the learning of logistic regression models using privileged information. We provide two new algorithms. For its development, the parameters of a conventional logistic regression trained with all available features, privileged and regular, are projected onto the parameter space associated to regular features (available at training and prediction time). The projection to obtain the model parameters is performed by the minimization of two different loss functions governed by logit terms and posterior probabilities. In addition, a metric is proposed to determine whether the use of privileged information can enhance performance. Experimental results report improvements of our proposals over the performance of conventional logistic regression learned without privileged information.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"615 ","pages":"Article 128869"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702527","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":"Cross-modal attention and geometric contextual aggregation network for 6DoF object pose estimation","authors":"Yi Guo ,&nbsp;Fei Wang ,&nbsp;Hao Chu ,&nbsp;Shiguang Wen","doi":"10.1016/j.neucom.2024.128891","DOIUrl":"10.1016/j.neucom.2024.128891","url":null,"abstract":"<div><div>The availability of affordable RGB-D sensors has made it more suitable to use RGB-D images for accurate 6D pose estimation, which allows for precise 6D parameter prediction using RGB-D images while maintaining a reasonable cost. A crucial research challenge is effectively exploiting adaptive feature extraction and fusion from the appearance information of RGB images and the geometric information of depth images. Moreover, previous methods have neglected the spatial geometric relationships of local position and the properties of point features, which are beneficial for tackling pose estimation in occlusion scenarios. In this work, we propose a cross-attention fusion framework for learning 6D pose estimation from RGB-D images. During the feature extraction stage, we design a geometry-aware context network that encodes local geometric properties of objects in point clouds using dual criteria, distance, and geometric angles. Moreover, we propose a cross-attention framework that combines spatial and channel attention in a cross-modal attention manner. This innovative framework enables us to capture the correlation and importance between RGB and depth features, resulting in improved accuracy in pose estimation, particularly in complex scenes. In the experimental results, we demonstrated that the proposed method outperforms state-of-the-art methods on four challenging benchmark datasets: YCB-Video, LineMOD, Occlusion LineMOD, and MP6D. Video is available at <span><span>https://youtu.be/4mgdbQKaHOc</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"617 ","pages":"Article 128891"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142745312","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":"Fully end-to-end EEG to speech translation using multi-scale optimized dual generative adversarial network with cycle-consistency loss","authors":"Chen Ma ,&nbsp;Yue Zhang ,&nbsp;Yina Guo ,&nbsp;Xin Liu ,&nbsp;Hong Shangguan ,&nbsp;Juan Wang ,&nbsp;Luqing Zhao","doi":"10.1016/j.neucom.2024.128916","DOIUrl":"10.1016/j.neucom.2024.128916","url":null,"abstract":"<div><div>Decoding auditory evoked electroencephalographic (EEG) signals to correlate them with speech acoustic features and construct transitional signals between different domain signals is a challenging and fascinating research topic. Brain–computer interface (BCI) technologies that incorporate auditory evoked potentials (AEPs) can not only leverage encoder–decoder architectures for signal decoding, but also employ generative adversarial networks (GANs) to translate from human neural activity to speech (T-HNAS). However, in previous research, the cascading ratio of transitional signals leads to varying degrees of information loss in the two-domain signals, and the optimal ratio of transitional signals differs across datasets, impacting the translation effectiveness. To address these issues, an improved dual generative adversarial network based on multi-scale optimization and cycle-consistency loss (MSCC-DualGAN) is proposed. We leverage the feature of cycle consistency loss, which facilitates cross-modal signal conversion, to replace transitional signals and maintain the integrity of signals in both domains during the loss computation process. Multi-scale optimization is utilized to refine the details of signals downsampled by the network, improving the similarity between features, thus enabling efficient, fully end-to-end EEG to speech translation. Furthermore, to validate the efficacy of this network, we construct a new EEG dataset and conduct studies using metrics such as mel cepstral distortion (MCD), pearson correlation coefficient (PCC), and structural similarity index measure (SSIM). Experimental results demonstrate that this new network significantly outperforms previous methods on auditory stimulus datasets.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128916"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142743067","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":"PSscheduler: A parameter synchronization scheduling algorithm for distributed machine learning in reconfigurable optical networks","authors":"Ling Liu ,&nbsp;Xiaoqiong Xu ,&nbsp;Pan Zhou ,&nbsp;Xi Chen ,&nbsp;Daji Ergu ,&nbsp;Hongfang Yu ,&nbsp;Gang Sun ,&nbsp;Mohsen Guizani","doi":"10.1016/j.neucom.2024.128876","DOIUrl":"10.1016/j.neucom.2024.128876","url":null,"abstract":"<div><div>With the increasing size of training datasets and models, parameter synchronization stage puts a heavy burden on the network, and communication has become one of the main performance bottlenecks of distributed machine learning (DML). Concurrently, optical circuit switch (OCS) with high bandwidth and reconfigurable features has increasingly introduced into the construction of network topology, obtaining the reconfigurable optical networks. Actually, OCS is conducive to accelerating the parameter synchronization stage, and thus improves training performance. However, unreasonable circuit scheduling algorithm has a great impact on parameter synchronization time because of non-negligible OCS switching delay. Besides, most of the existing circuit scheduling algorithms do not effectively use the training characteristics of DML, and the performance gains are limited. Therefore, in this paper, we study the parameter synchronization scheduling algorithm in reconfigurable optical networks, and propose PSscheduler by jointly optimizing the circuit scheduling and deployment of parameter servers in parameter server (PS) architecture. Specifically, a mathematical optimization model is established first, which takes into account the deployment of parameter servers, the allocation of parameter blocks and circuit scheduling. Subsequently, the mathematical model is solved by relaxed variables and deterministic rounding approach. The results of simulation based on real DML workloads demonstrate that compared to <em>Sunflow</em> and <em>HLF</em> , PSscheduler is more stable and can reduce parameter synchronization time (PST) by up to 46.61% and 25%, respectively.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128876"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142743063","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":"Potential Knowledge Extraction Network for Class-Incremental Learning","authors":"Xidong Xi ,&nbsp;Guitao Cao ,&nbsp;Wenming Cao ,&nbsp;Yong Liu ,&nbsp;Yan Li ,&nbsp;Hong Wang ,&nbsp;He Ren","doi":"10.1016/j.neucom.2024.128923","DOIUrl":"10.1016/j.neucom.2024.128923","url":null,"abstract":"<div><div>Class-Incremental Learning (CIL) aims to dynamically learn new classes without forgetting the old ones, and it is typically achieved by extracting knowledge from old data and continuously transferring it to new tasks. In the replay-based approaches, selecting appropriate exemplars is of great importance since exemplars represent the most direct form of retaining old knowledge. In this paper, we propose a novel CIL framework: <em>Potential Knowledge Extraction Network</em> (PKENet), which addresses the issue of neglecting the knowledge of inter-sample relation in most existing works and suggests an innovative approach for exemplar selection. Specifically, to address the challenge of knowledge transfer, we design a <em>relation consistency loss</em> and a <em>hybrid cross-entropy loss</em>, where the former works by extracting structural knowledge from the old model while the latter captures graph-wise knowledge, enabling the new model to acquire more old knowledge. To enhance the anti-forgetting effect of exemplar set, we devise a <em>maximum-forgetting-priority</em> method for selecting samples most susceptible to interference from the model’s update. To overcome the prediction bias problem in CIL, we introduce the Total Direct Effect inference method into our model. Experimental results on CIFAR100, ImageNet-Full and ImageNet-Subset datasets show that multiple state-of-the-art CIL methods can be directly combined with our PKENet to reap significant performance improvement. Code: <span><span>https://github.com/XXDyeah/PKENet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128923"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142742961","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":"Superpixel semantics representation and pre-training for vision–language tasks","authors":"Siyu Zhang ,&nbsp;Yeming Chen ,&nbsp;Yaoru Sun ,&nbsp;Fang Wang ,&nbsp;Jun Yang ,&nbsp;Lizhi Bai ,&nbsp;Shangce Gao","doi":"10.1016/j.neucom.2024.128895","DOIUrl":"10.1016/j.neucom.2024.128895","url":null,"abstract":"<div><div>The key to integrating visual language tasks is to establish a good alignment strategy. Recently, visual semantic representation has achieved fine-grained visual understanding by dividing grids or image patches. However, the coarse-grained semantic interactions in image space should not be ignored, which hinders the extraction of complex contextual semantic relations at the scene boundaries. This paper proposes superpixels as comprehensive and robust visual primitives, which mine coarse-grained semantic interactions by clustering perceptually similar pixels, speeding up the subsequent processing of primitives. To capture superpixel-level semantic features, we propose a Multiscale Difference Graph Convolutional Network (MDGCN). It allows parsing the entire image as a fine-to-coarse visual hierarchy. To reason actual semantic relations, we reduce potential noise interference by aggregating difference information between adjacent graph nodes. Finally, we propose a multi-level fusion rule in a bottom-up manner to avoid understanding deviation by mining complementary spatial information at different levels. Experiments show that the proposed method can effectively promote the learning of multiple downstream tasks. Encouragingly, our method outperforms previous methods on all metrics.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"615 ","pages":"Article 128895"},"PeriodicalIF":5.5,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142703004","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":"The effect of the head number for multi-head self-attention in remaining useful life prediction of rolling bearing and interpretability","authors":"Qiwu Zhao,&nbsp;Xiaoli Zhang,&nbsp;Fangzhen Wang,&nbsp;Panfeng Fan,&nbsp;Erick Mbeka","doi":"10.1016/j.neucom.2024.128946","DOIUrl":"10.1016/j.neucom.2024.128946","url":null,"abstract":"<div><div>As one of the machine learning (ML) models, the multi-head self-attention mechanism (MSM) is competent in encoding high-level feature representations, providing computing superiorities, and systematically processing sequences bypassing the recurrent neural networks (RNN) models. However, the model performance and computational results are affected by head number, and the lack of impact interpretability has become a primary obstacle due to the complex internal working mechanisms. Therefore, the effects of the head number of the MSM on the accuracy of the result, the robustness of the model, and computation efficiency are investigated in the remaining useful life (RUL) prediction of rolling bearings. The results show that the accuracy of prediction results will be reduced caused by large or few head numbers. In addition, the more heads are selected, the more robust and higher the predictive efficiency of the model is achieved. The above effects are explained relying on the visualization of the attention weight distribution and functional networks, which are constructed and solved by the equivalent fully connected layer and graph theory analysis, respectively. The model's attention coefficient distribution during training and prediction shows that the representative information will be captured inadequately if fewer heads are selected, which causes MSM to neglect to assign large attention coefficients to degraded information. On the contrary, representational degradation information and redundant information will be acquired by models with too many heads. MSM will be disturbed by this redundant information in the attention weight distribution, resulting in incorrect allocation of attention. Both of these cases will reduce the accuracy of the prediction results. In addition, the selection rules of the head number are established based on the feature complexity that is measured by the sample entropy (SamEn). The local range for head selection is also found based on the relationship between head number and feature complexity; The effects of the head number of the MSM on the robustness of the model and computation efficiency are explained by the changes in the three parameters (average of the clustering coefficients, global efficiency, and of the average shortest path length) of the graph, which is constructed after solving the function network. The research provides a reference for rolling bearing prediction with high computational accuracy, calculation efficiency, and strong robustness using MSM.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128946"},"PeriodicalIF":5.5,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142742962","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":"CRMSP: A semi-supervised approach for key information extraction with Class-Rebalancing and Merged Semantic Pseudo-Labeling","authors":"Qi Zhang,&nbsp;Yonghong Song,&nbsp;Pengcheng Guo,&nbsp;Yangyang Hui","doi":"10.1016/j.neucom.2024.128907","DOIUrl":"10.1016/j.neucom.2024.128907","url":null,"abstract":"<div><div>There is a growing demand in the field of Key Information Extraction (KIE) to apply semi-supervised learning (SSL) to save manpower and costs, as training document data using fully-supervised methods requires labor-intensive manual annotation. The main challenges of applying SSL in the KIE are (1) underestimation of the confidence of tail classes in the long-tailed distribution and (2) difficulty in achieving intra-class compactness and inter-class separability of tail features. To address these challenges, we propose a novel semi-supervised approach for KIE with Class-Rebalancing and Merged Semantic Pseudo-Labeling (CRMSP). Firstly, the Class-Rebalancing Pseudo-Labeling (CRP) module introduces a reweighting factor to rebalance pseudo-labels, increasing attention to tail classes. Secondly, we propose the Merged Semantic Pseudo-Labeling (MSP) module to cluster tail features of unlabeled data by assigning samples to Merged Prototypes (MP). Additionally, we designed a new contrastive loss specifically for MSP. Extensive experimental results on three well-known benchmarks demonstrate that CRMSP achieves state-of-the-art performance. Remarkably, CRMSP achieves 3.24% f1-score improvement over state-of-the-art on the CORD.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128907"},"PeriodicalIF":5.5,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142742974","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":"HifiDiff: High-fidelity diffusion model for face hallucination from tiny non-frontal faces","authors":"Wei Wang,&nbsp;Xing Wang,&nbsp;Yuguang Shi,&nbsp;Xiaobo Lu","doi":"10.1016/j.neucom.2024.128882","DOIUrl":"10.1016/j.neucom.2024.128882","url":null,"abstract":"<div><div>Obtaining a high-quality frontal facial image from a low-resolution (LR) non-frontal facial image is crucial for many facial analysis tasks. Recently, diffusion models (DMs) have made impressive progress in near-frontal face super-resolution. However, when faced with non-frontal LR faces, the existing DMs exhibit poor identity preservation and facial detail fidelity. In this paper, we present a novel high-fidelity DM named HifiDiff for simultaneously super-resolving and frontalizing tiny non-frontal facial images. It consists of a two-stage pipeline: facial preview and facial refinement. In the first stage, we pretrain a coarse restoration module to obtain a coarse high-resolution (HR) frontal face, which serves as a superior constraint condition to enhance the ability to solve complex inverse transform issues. In the second stage, we leverage the strong generation capabilities of the latent DM to refine the facial details. Specifically, we design a two-pathway control structure that consists of a facial prior guidance (FPG) module and an identity consistency (IDC) module to control the denoising process. FPG encodes multilevel features derived from latent coarse HR frontal faces and employs hybrid cross-attention to capture their intrinsic correlations with the denoiser features, thereby improving the fidelity of the facial details. IDC utilizes contrastive learning to extract high-level semantic identity-representing features to constrain the denoiser, thereby maintaining the fidelity of facial identities. Extensive experiments demonstrate that our HifiDiff produces both high-fidelity and realistic HR frontal facial images, surpassing other state-of-the-art methods in qualitative and quantitative analyses, as well as in downstream facial recognition tasks.</div></div>","PeriodicalId":19268,"journal":{"name":"Neurocomputing","volume":"616 ","pages":"Article 128882"},"PeriodicalIF":5.5,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142742971","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}