Neural Networks最新文献

{"title":"SPAST: Arbitrary style transfer with style priors via pre-trained large-scale model","authors":"Zhanjie Zhang ,&nbsp;Quanwei Zhang ,&nbsp;Junsheng Luan ,&nbsp;Mengyuan Yang ,&nbsp;Yun Wang ,&nbsp;Lei Zhao","doi":"10.1016/j.neunet.2025.107556","DOIUrl":"10.1016/j.neunet.2025.107556","url":null,"abstract":"<div><div>Given an arbitrary content and style image, arbitrary style transfer aims to render a new stylized image which preserves the content image’s structure and possesses the style image’s style. Existing arbitrary style transfer methods are based on either small models or pre-trained large-scale models. The small model-based methods fail to generate high-quality stylized images, bringing artifacts and disharmonious patterns. The pre-trained large-scale model-based methods can generate high-quality stylized images but struggle to preserve the content structure and cost long inference time. To this end, we propose a new framework, called SPAST, to generate high-quality stylized images with less inference time. Specifically, we design a novel Local–global Window Size Stylization Module (LGWSSM) to fuse style features into content features. Besides, we introduce a novel style prior loss, which can dig out the style priors from a pre-trained large-scale model into the SPAST and motivate the SPAST to generate high-quality stylized images with short inference time. We conduct abundant experiments to verify that our proposed method can generate high-quality stylized images and less inference time compared with the SOTA arbitrary style transfer methods.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107556"},"PeriodicalIF":6.0,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068499","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":"Modeling structured data learning with Restricted Boltzmann machines in the teacher–student setting","authors":"Robin Thériault ,&nbsp;Francesco Tosello ,&nbsp;Daniele Tantari","doi":"10.1016/j.neunet.2025.107542","DOIUrl":"10.1016/j.neunet.2025.107542","url":null,"abstract":"<div><div>Restricted Boltzmann machines (RBM) are generative models capable to learn data with a rich underlying structure. We study the teacher–student setting where a student RBM learns structured data generated by a teacher RBM. The amount of structure in the data is controlled by adjusting the number of hidden units of the teacher and the correlations in the rows of the weights, a.k.a. patterns. In the absence of correlations, we validate the conjecture that the performance is independent of the number of teacher patterns and hidden units of the student RBMs, and we argue that the teacher–student setting can be used as a toy model for studying the lottery ticket hypothesis. Beyond this regime, we find that the critical amount of data required to learn the teacher patterns decreases with both their number and correlations. In both regimes, we find that, even with a relatively large dataset, it becomes impossible to learn the teacher patterns if the inference temperature used for regularization is kept too low. In our framework, the student can learn teacher patterns one-to-one or many-to-one, generalizing previous findings about the teacher–student setting with two hidden units to any arbitrary finite number of hidden units.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107542"},"PeriodicalIF":6.0,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144084305","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":"SurvGraph: A hybrid-graph attention network for survival prediction using whole slide pathological images in gastric cancer","authors":"Yuanshen Zhao ,&nbsp;Longsong Li ,&nbsp;Xi Yu ,&nbsp;Ke Han ,&nbsp;Jingxian Duan ,&nbsp;Dong Liang ,&nbsp;Ningli Chai ,&nbsp;Zhi-Cheng Li","doi":"10.1016/j.neunet.2025.107607","DOIUrl":"10.1016/j.neunet.2025.107607","url":null,"abstract":"<div><div>Whole slide pathological images have shown significant potential for patient prognostication. Graph representation learning provides a robust framework for in-depth analysis of whole-slide images to construct predictive models. In this study, we introduce SurvGraph, an innovative graph-based deep learning network designed for gastric cancer survival prediction using whole slide pathological images. SurvGraph employs a hybrid graph construction approach that integrates multiple feature types, including color, texture, and deep learning features extracted from the pathological images to build node representations. SurvGraph utilizes a multi-head attention graph network, which performs survival prediction based on the graph structure. We evaluate the SurvGraph model on a large dataset of 708 gastric cancer patients from three independent cohorts for overall survival prediction. To assess the impact of various feature sets, we examine their performance when used individually and in combination. With five-fold cross-validation, our results demonstrate that the SurvGraph model achieves an average concordance index (C-index) of 0.706 with a standard deviation (SD) of 0.019. The proposed SurvGraph model has also attained a C-index of 0.708 (SD = 0.040) in the external testing set. In addition to baseline comparisons, we conducted a comprehensive benchmarking study comparing SurvGraph against established graph neural network architectures and multiple instance learning-based deep learning frameworks. The results indicate that the SurvGraph model outperforms the compared prediction models, suggesting its potential as a valuable tool for enhancing gastric cancer prognosis estimation.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107607"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143943710","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":"Wp-VTON: A wrinkle-preserving virtual try-on network via clothing texture book","authors":"Xiangyu Mu ,&nbsp;Haijun Zhang ,&nbsp;Jianghong Ma ,&nbsp;Zhao Zhang ,&nbsp;Lin Jiang ,&nbsp;Xiao Chen ,&nbsp;Feng Jiang","doi":"10.1016/j.neunet.2025.107546","DOIUrl":"10.1016/j.neunet.2025.107546","url":null,"abstract":"<div><div>Virtual try-on technology seeks to seamlessly integrate an image of a specified garment onto the target person, generating a synthesized image that realistically depicts the person wearing the clothing. Existing methods based on generative adversarial network (GAN) for clothing warping in the generation process usually use human pose- and body parsing-based features to guide the distortion of a flattened clothing item. However, it is hard using these approaches to accurately capture the spatial characteristics of the distorted clothing (e.g., wrinkles on the clothing). In this research, we propose a Wrinkle-Preserving Virtual Try-On Network, named WP-VTON, to address the aforementioned issues exist in the virtual try-on task. Specifically, in the clothing warping stage, we incorporate the normal features extracted from spatial attributes of both clothing and the human body to learn about the clothing deformation caused by warping; in the try-on generation stage, we leverage a pre-trained StyleGAN, called clothing texture book, to optimize the try-on image, with the aim of further improving the generation capability of WP-VTON with regard to texture details. Experimental results in public datasets demonstrate the effectiveness of our method by outperforming the state-of-the-art GAN-based virtual try-on models.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107546"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143934692","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":"Multiple-input and multiple-output encoders with DNA-based winner-take-all neural Networks","authors":"Chun Huang,&nbsp;Qingshuang Guo,&nbsp;Jiaying Shao,&nbsp;Baolei Peng,&nbsp;Panlong Li,&nbsp;Junwei Sun,&nbsp;Yanfeng Wang","doi":"10.1016/j.neunet.2025.107487","DOIUrl":"10.1016/j.neunet.2025.107487","url":null,"abstract":"<div><div>DNA logic circuits are essential building blocks for molecular computers. Traditional molecular logic circuits primarily use basic gate circuits as computational units, achieving complex functions via multiple cascades. However, even simple logical functions often require complex cascading processes. This study introduces Winner-take-all (WTA) neural networks based on DNA strand displacement, harnessing neural networks’ powerful computational capabilities for solving nonlinear complex problems. We developed multifunctional encoder circuits for two-bit and three-bit outputs and extended this design into a universal encoder circuit model. Furthermore, by cascading two DNA WTA neural networks, we successfully constructed a two-layer neural network that implements a four-bit priority encoder circuit. Simulations were performed and validated using Visual DSD software. Experimental results reveal the significant potential of DNA neural networks for building ultra-large-scale molecular logic circuits. This study offers fresh perspectives on the functionality of DNA neural networks and proposes a novel methodology for constructing complex molecular logic circuits.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107487"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068396","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":"Improved two-view interactional fuzzy learning based on mutual-rectification and knowledge-mergence","authors":"Ta Zhou ,&nbsp;Wei Yan ,&nbsp;Zhengxin Xia ,&nbsp;Shuihua Wang ,&nbsp;Ge Ren ,&nbsp;Bing Li ,&nbsp;Weiping Ding ,&nbsp;Jing Cai","doi":"10.1016/j.neunet.2025.107576","DOIUrl":"10.1016/j.neunet.2025.107576","url":null,"abstract":"<div><div>Nasopharyngeal carcinoma (NPC) is a malignant tumor that originates from the back of the nasal canal from above the soft palate to the upper larynx. Because the nasopharyngeal location is deeply hidden, it is often difficult for a single imaging means to clarify its complex adjacency. In addition, there exist some differences and uncertainties in its clinical manifestations. Although two-view fuzzy classifiers can effectively tap into the nasopharyngeal location for hidden information and exhibit good classification performance, existing fuzzy reasoning for predicting whether or not a nasopharyngeal cancer often stems from the inability to reuse the one-sided rules. Therefore, a novel two-view mutual rectification and knowledge mergence Takagi-Sugeno-Kang fuzzy classifier (TVRM-TFC) is proposed here to address the challenge of using imaging means to fine-tune the organ tissues. Firstly, Kullback-Leibler divergence (KLIC) is used to select important features from various imaging sections (i.e., pieces of knowledge). Secondly, the interpretable zero-order Takagi-Sugeno-Kang (TSK) fuzzy classifier is used as the basic training unit to simultaneously obtain satisfactory accuracies and concise linguistic interpretability. Thirdly, from the perspective of both imaging means and the organ, this study fine-tunes the information required for decision-making between different imaging means, so that the complementary advantages of the different views may improve the decision-making information and thus increase decision accuracies. Finally, the perspective of imaging technology and the organ are merged to capture decision-making knowledge. These decision-making advantages from different views are organically integrated to compensate information and further optimize the decision-making information. The merits of the proposed classifier are demonstrated through comparative experimental analysis on CT and MRI data.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107576"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143929465","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":"Pruning the ensemble of convolutional neural networks using second-order cone programming","authors":"Buse Çisil Güldoğuş ,&nbsp;Abdullah Nazhat Abdullah ,&nbsp;Muhammad Ammar Ali ,&nbsp;Süreyya Özöǧür-Akyüz","doi":"10.1016/j.neunet.2025.107544","DOIUrl":"10.1016/j.neunet.2025.107544","url":null,"abstract":"<div><div>Ensemble techniques are frequently encountered in machine learning and engineering problems since the method combines different models and produces an optimal predictive solution. The ensemble concept can be adapted to deep learning models to provide robustness and reliability. Due to the growth of the models in deep learning, using ensemble pruning is highly important to deal with computational complexity. Hence, this study proposes a mathematical model which prunes the ensemble of Convolutional Neural Networks (CNNs) consisting of different depths and layers that maximizes accuracy and diversity simultaneously with a sparse second order conic optimization model. The proposed model is tested on the CIFAR-10, CIFAR-100, and MNIST datasets, and its performance is compared with benchmark pruning methods, yielding promising results while reducing model complexity.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107544"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143943615","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":"Modeling multi-scale uncertainty with evidence integration for reliable polyp segmentation","authors":"Xiaolu Kang ,&nbsp;Zhuoqi Ma ,&nbsp;Kang Liu,&nbsp;Yunan Li,&nbsp;Qiguang Miao","doi":"10.1016/j.neunet.2025.107553","DOIUrl":"10.1016/j.neunet.2025.107553","url":null,"abstract":"<div><div>Polyp segmentation is critical in medical image analysis. Traditional methods, while capable of producing precise outputs in well-defined regions, often struggle with blurry or ambiguous areas in medical images, which can lead to errors in clinical decision-making. Additionally, these methods typically generate only a single deterministic segmentation result, failing to account for the inherent uncertainty in the segmentation process. This limitation undermines the reliability of segmentation models in clinical practice, as they lack the ability to provide insights into the confidence or certainty of their predictions, leaving clinicians skeptical of their utility. To address these challenges, we propose a novel multi-scale uncertainty modeling framework for polyp segmentation, grounded in evidence theory. Our approach leverages the Dirichlet distribution to classify pixels within polyp images while integrating uncertainty across different scales. We first employ an Uncertainty Region Enhancement Process (UREP) to refine uncertain regions and Integrated Balance Module (IBM) to dynamically balance the weights between different feature maps for generating semantic fusion feature maps. Subsequently, we utilize two feature extraction sub-networks to learn feature representations from original images and semantic fusion feature maps. We further develop a Multi-scale Evidence Integration Network (MEIN) to robustly model uncertainty through subjective logic, merging results from two sub-networks to ensure a comprehensive understanding of uncertainty and produce reliable segmentation results. In contrast to most existing methods, our approach not only generates segmentation results but also provides uncertainty estimates, offering clinicians valuable insights into the reliability of the predictions. Experimental results on five polyp segmentation datasets demonstrate that our proposed method remains competitive and generates effective uncertainty estimations compared to existing representative methods. The code is available at <span><span>https://github.com/q1216355254/MEIN</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107553"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144115683","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":"Implementing feature binding through dendritic networks of a single neuron","authors":"Yuanhong Tang ,&nbsp;Shanshan Jia ,&nbsp;Tiejun Huang ,&nbsp;Zhaofei Yu ,&nbsp;Jian K. Liu","doi":"10.1016/j.neunet.2025.107555","DOIUrl":"10.1016/j.neunet.2025.107555","url":null,"abstract":"<div><div>A single neuron receives an extensive array of synaptic inputs through its dendrites, raising the fundamental question of how these inputs undergo integration and summation, culminating in the initiation of spikes in the soma. Experimental and computational investigations have revealed various modes of integration operations that include linear, superlinear, and sublinear summation. Interestingly, different types of neurons exhibit diverse patterns of dendritic integration depending on the spatial distribution of dendrites. The functional implications of these specific integration modalities remain largely unexplored. In this study, we employ the Purkinje cell (PC) as a model system to investigate these complex questions. Our findings reveal that PCs generally exhibit sublinear summation across their expansive dendrites. Both spatial and temporal input dynamically modulates the degree of sublinearity. Strong sublinearity necessitates the synaptic distribution in PCs to be globally scattered sensitive, whereas weak sublinearity facilitates the generation of complex firing patterns in PCs. Using dendritic branches characterized by strong sublinearity as computational units, we demonstrate that a neuron can successfully address the feature binding problem. Taken together, these results offer a systematic perspective on the functional role of dendritic sublinearity, inspiring a broader understanding of dendritic integration in various neuronal types.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107555"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143947842","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":"Enhancing the transferability of adversarial attacks via Scale Enriching","authors":"Yuhang Zhao ,&nbsp;Jun Zheng ,&nbsp;Xianfeng Gao ,&nbsp;Lu Liu ,&nbsp;Yaoyuan Zhang ,&nbsp;Quanxin Zhang","doi":"10.1016/j.neunet.2025.107549","DOIUrl":"10.1016/j.neunet.2025.107549","url":null,"abstract":"<div><div>Deep learning models are vulnerable to adversarial attacks. Transfer-based adversarial examples are crafted against surrogate models and transferred to victim models. However, under the black-box settings, most adversaries have poor transferability on models with different input sizes. In this work, we propose the Scale Enriching Method (SEM) to enhance the transferability of adversarial examples using an input scale-enriching framework. By scaling the surrogate model’s input in a specific range, our method enriches the attention areas that the surrogate model perceives and enlarges the tolerance of the distinction among different models, significantly improving the transferability. Notably, SEM avoids introducing extraneous noise during perturbation generation, thereby preserving the inherent textural features corresponding to different scales within the original images. Experiments on ImageNet show that our method successfully mitigates the gap of transferability between models with different input sizes. Furthermore, we demonstrate that our method can integrate with existing methods and bypass a variety of defense methods with over 90% success rate.</div></div>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"189 ","pages":"Article 107549"},"PeriodicalIF":6.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068383","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}