Information Fusion最新文献

{"title":"Applications of knowledge distillation in remote sensing: A survey","authors":"Yassine Himeur ,&nbsp;Nour Aburaed ,&nbsp;Omar Elharrouss ,&nbsp;Iraklis Varlamis ,&nbsp;Shadi Atalla ,&nbsp;Wathiq Mansoor ,&nbsp;Hussain Al-Ahmad","doi":"10.1016/j.inffus.2024.102742","DOIUrl":"10.1016/j.inffus.2024.102742","url":null,"abstract":"<div><div>With the ever-growing complexity of models in the field of remote sensing (RS), there is an increasing demand for solutions that balance model accuracy with computational efficiency. Knowledge distillation (KD) has emerged as a powerful tool to meet this need, enabling the transfer of knowledge from large, complex models to smaller, more efficient ones without significant loss in performance. This review article provides an extensive examination of KD and its innovative applications in RS. KD, a technique developed to transfer knowledge from a complex, often cumbersome model (teacher) to a more compact and efficient model (student), has seen significant evolution and application across various domains. Initially, we introduce the fundamental concepts and historical progression of KD methods. The advantages of employing KD are highlighted, particularly in terms of model compression, enhanced computational efficiency, and improved performance, which are pivotal for practical deployments in RS scenarios. The article provides a comprehensive taxonomy of KD techniques, where each category is critically analyzed to demonstrate the breadth and depth of the alternative options, and illustrates specific case studies that showcase the practical implementation of KD methods in RS tasks, such as instance segmentation and object detection. Further, the review discusses the challenges and limitations of KD in RS, including practical constraints and prospective future directions, providing a comprehensive overview for researchers and practitioners in the field of RS. Through this organization, the paper not only elucidates the current state of research in KD but also sets the stage for future research opportunities, thereby contributing significantly to both academic research and real-world applications.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102742"},"PeriodicalIF":14.7,"publicationDate":"2024-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142554838","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":"Label distribution-driven multi-view representation learning","authors":"Wenbiao Yan ,&nbsp;Minghong Wu ,&nbsp;Yiyang Zhou ,&nbsp;Qinghai Zheng ,&nbsp;Jinqian Chen ,&nbsp;Haozhe Cheng ,&nbsp;Jihua Zhu","doi":"10.1016/j.inffus.2024.102727","DOIUrl":"10.1016/j.inffus.2024.102727","url":null,"abstract":"<div><div>In multi-view representation learning (MVRL), the challenge of category uncertainty is significant. Existing methods excel at deriving shared representations across multiple views, but often neglect the uncertainty associated with cluster assignments from each view, thereby leading to increased ambiguity in the category determination. Additionally, methods like kernel-based or neural network-based approaches, while revealing nonlinear relationships, lack attention to category uncertainty. To address these limitations, this paper proposes a method leveraging the uncertainty of label distributions to enhance MVRL. Specifically, our approach combines uncertainty reduction based on label distribution with view representation learning to improve clustering accuracy and robustness. It initially computes the within-view representation of the sample and semantic labels. Then, we introduce a novel constraint based on either variance or information entropy to mitigate class uncertainty, thereby improving the discriminative power of the learned representations. Extensive experiments conducted on diverse multi-view datasets demonstrate that our method consistently outperforms existing approaches, producing more accurate and reliable class assignments. The experimental results highlight the effectiveness of our method in enhancing MVRL by reducing category uncertainty and improving overall classification performance. This method is not only very interpretable but also enhances the model’s ability to learn multi-view consistent information.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102727"},"PeriodicalIF":14.7,"publicationDate":"2024-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142554834","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":"HSMix: Hard and soft mixing data augmentation for medical image segmentation","authors":"D. Sun ,&nbsp;F. Dornaika ,&nbsp;N. Barrena","doi":"10.1016/j.inffus.2024.102741","DOIUrl":"10.1016/j.inffus.2024.102741","url":null,"abstract":"<div><div>Due to the high cost of annotation or the rarity of some diseases, medical image segmentation is often limited by data scarcity and the resulting overfitting problem. Self-supervised learning and semi-supervised learning can mitigate the data scarcity challenge to some extent. However, both of these paradigms are complex and require either hand-crafted pretexts or well-defined pseudo-labels. In contrast, data augmentation represents a relatively simple and straightforward approach to addressing data scarcity issues. It has led to significant improvements in image recognition tasks. However, the effectiveness of local image editing augmentation techniques in the context of segmentation has been less explored. Additionally, traditional data augmentation methods for local image editing augmentation methods generally utilize square regions, which cause a loss of contour information.</div><div>We propose HSMix, a novel approach to local image editing data augmentation involving hard and soft mixing for medical semantic segmentation. In our approach, a hard-augmented image is created by combining homogeneous regions (superpixels) from two source images. A soft mixing method further adjusts the brightness of these composed regions with brightness mixing based on locally aggregated pixel-wise saliency coefficients. The ground-truth segmentation masks of the two source images undergo the same mixing operations to generate the associated masks for the augmented images.</div><div>Our method fully exploits both the prior contour and saliency information, thus preserving local semantic information in the augmented images while enriching the augmentation space with more diversity. Our method is a plug-and-play solution that is model agnostic and applicable to a range of medical imaging modalities. Extensive experimental evidence has demonstrated its effectiveness in a variety of medical segmentation tasks. The source code is available in <span><span>https://github.com/DanielaPlusPlus/HSMix</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102741"},"PeriodicalIF":14.7,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An adaptive meta-imitation learning-based recommendation environment simulator: A case study on ship-cargo matching","authors":"Guangyao Pang ,&nbsp;Jiehang Xie ,&nbsp;Fei Hao","doi":"10.1016/j.inffus.2024.102740","DOIUrl":"10.1016/j.inffus.2024.102740","url":null,"abstract":"<div><div>High-quality shipping is one of the effective ways for sustainable cities in inland river basins to improve transportation efficiency and reduce energy consumption. Currently, the biggest challenge faced by shipping is the high empty-ship rate, which makes it impossible to directly apply machine learning methods due to the cold-start problem. Although some researchers have tried to utilize deep reinforcement learning(DRL)-based recommendation that do not rely on manually labeled data to alleviate the cold-start problem, progress has been slow due to the lack of available training environment. Therefore, this paper introduces an adaptive meta-imitation learning-based recommendation environment simulator, termed AMIL-Simulator. Specifically, we construct a conditionally guided diffusion model to simulate shipowner behavior in a dynamically changing environment. Moreover, we propose a shipowner reward model based on adaptive meta-imitation learning, enabling the learning of shipowner rewards across multiple tasks, even when confronted with limited samples and imbalanced categories. By conducting extensive quantitative experimental evaluations and shipowner-cargo matching studies, the results demonstrate the effectiveness of AMIL-Simulator, particularly in smaller-scale and cold-start environments.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102740"},"PeriodicalIF":14.7,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531896","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":"Recent advances in data-driven fusion of multi-modal imaging and genomics for precision medicine","authors":"Shuo Wang ,&nbsp;Meng Liu ,&nbsp;Yan Li ,&nbsp;Xinyu Zhang ,&nbsp;Mengting Sun ,&nbsp;Zian Wang ,&nbsp;Ruokun Li ,&nbsp;Qirong Li ,&nbsp;Qing Li ,&nbsp;Yili He ,&nbsp;Xumei Hu ,&nbsp;Longyu Sun ,&nbsp;Fuhua Yan ,&nbsp;Mengyao Yu ,&nbsp;Weiping Ding ,&nbsp;Chengyan Wang","doi":"10.1016/j.inffus.2024.102738","DOIUrl":"10.1016/j.inffus.2024.102738","url":null,"abstract":"<div><div>Imaging genomics is poised to revolutionize clinical practice by providing deep insights into the genetic underpinnings of disease, enabling early detection, and facilitating personalized treatment strategies. The field has seen remarkable advancements, with significant momentum fueled by cutting-edge imaging techniques, sophisticated data-driven fusion methods, and extensive large cohort datasets. Originally centered on the brain, imaging genomics has now expanded to encompass other organs throughout the body. Due to the highly interdisciplinary nature involving medical imaging, genetics, machine learning, and clinical medicine, readers who wish to conduct research in this field urgently need a comprehensive review. This survey provides an overview of recent advancements in data-driven fusion of multi-modal imaging and genomics, covering applications in the brain, heart, lungs, breasts, abdomen, and bones. We summarize three primary fusion strategies: correlation analysis, causal analysis, and machine learning, discussing their respective application scenarios. Additionally, we explore clinical applications that integrate imaging datasets and genomic data across six major organ systems, and present available open datasets featuring both modalities. Finally, we summarize the challenges and future directions in imaging genomics, which include improving data representation, integrating other omics data, conducting cross-dataset analyses, advancing machine learning algorithms, and investigating organ interactions. This survey aims to review the latest developments in data-driven fusion for precision medicine while providing insights into the future of this evolving field.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102738"},"PeriodicalIF":14.7,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531902","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":"An interactive iteration consensus based social network large-scale group decision making method and its application in zero-waste city evaluation","authors":"Fanyong Meng ,&nbsp;Hao Li ,&nbsp;Jinyu Li","doi":"10.1016/j.inffus.2024.102744","DOIUrl":"10.1016/j.inffus.2024.102744","url":null,"abstract":"<div><div>The construction of zero-waste (ZW) cities receives increasing attention from the Chinese government. The evaluation is essential to make policy variations according to the actual situation in each place. Previous assessments of ZW cities have primarily relied on historical data, which fails to account for the subjective preferences of various stakeholders. For example, it is challenging to capture residents' subjective opinions about the development of a ZW city. This paper presents a social network large-scale group decision-making method for evaluating the construction of ZW city. First, experts' evaluation opinions and trust relations are used to develop an improved clustering method. The weights of the clusters are then determined using internal-external cohesion indices and the number of experts, with experts' weights defined by their similarity-trust degree. An optimization model based on interactive iteration consensus is formulated, considering the fairness and rationality of allocation schemes. Additionally, a new social network large-scale group decision-making method is presented. Finally, the proposed method is illustrated with a case study of selecting a national-level ZW city in Jiangsu Province.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102744"},"PeriodicalIF":14.7,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531897","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":"Multi-view support vector machine classifier via L0/1 soft-margin loss with structural information","authors":"Chen Chen ,&nbsp;Qianfei Liu ,&nbsp;Renpeng Xu ,&nbsp;Ying Zhang ,&nbsp;Huiru Wang ,&nbsp;Qingmin Yu","doi":"10.1016/j.inffus.2024.102733","DOIUrl":"10.1016/j.inffus.2024.102733","url":null,"abstract":"<div><div>Multi-view learning seeks to leverage the advantages of various views to complement each other and make full use of the latent information in the data. Nevertheless, effectively exploring and utilizing common and complementary information across diverse views remains challenging. In this paper, we propose two multi-view classifiers: multi-view support vector machine via <span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></math></span> soft-margin loss (Mv<span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></math></span>-SVM) and structural Mv<span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></math></span>-SVM (Mv<span><math><mrow><mi>S</mi><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></mrow></math></span>-SVM). The key difference between them is that Mv<span><math><mrow><mi>S</mi><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></mrow></math></span>-SVM additionally fuses structural information, which simultaneously satisfies the consensus and complementarity principles. Despite the discrete nature inherent in the <span><math><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></math></span> soft-margin loss, we successfully establish the optimality theory for Mv<span><math><mrow><mi>S</mi><msub><mrow><mi>L</mi></mrow><mrow><mn>0</mn><mo>/</mo><mn>1</mn></mrow></msub></mrow></math></span>-SVM. This includes demonstrating the existence of optimal solutions and elucidating their relationships with P-stationary points. Drawing inspiration from the P-stationary point optimality condition, we design and integrate a working set strategy into the proximal alternating direction method of multipliers. This integration significantly enhances the overall computational speed and diminishes the number of support vectors. Last but not least, numerical experiments show that our suggested models perform exceptionally well and have faster computational speed, affirming the rationality and effectiveness of our methods.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102733"},"PeriodicalIF":14.7,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531901","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":"Open knowledge graph completion with negative-aware representation learning and multi-source reliability inference","authors":"Huang Peng,&nbsp;Weixin Zeng,&nbsp;Jiuyang Tang,&nbsp;Mao Wang,&nbsp;Hongbin Huang,&nbsp;Xiang Zhao","doi":"10.1016/j.inffus.2024.102729","DOIUrl":"10.1016/j.inffus.2024.102729","url":null,"abstract":"<div><div>Multi-source data fusion is essential for building smart cities by providing a comprehensive and holistic understanding of urban environments. Specifically, smart city-oriented knowledge graphs (KGs) require supplementary information from other open sources to increase their completeness, thus better supporting downstream tasks for smart cities. Nevertheless, existing open knowledge graph completion (KGC) approaches often overlook source quality assessment and fail to fully utilize prior knowledge, which tend to yield less satisfying results. To fill in these gaps, in this work, we propose a new open KGC method with negative-aware representation learning and multi-source reliability inference, i.e., <span>Nari</span>, which can effectively integrate the multi-source data concerning sustainable cities, providing reliable knowledge for downstream tasks. Specifically, we first train a graph neural network based encoder with a novel negative sampling strategy to better characterize prior knowledge in KG, and then identify new facts based on the learned prior knowledge and source reliability. The experiments on both general benchmark and waterlogging benchmark pertaining to sustainable cities demonstrate the effectiveness and wide applicability of <span>Nari</span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102729"},"PeriodicalIF":14.7,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531892","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":"Blockchain-based privacy-preserving incentive scheme for internet of electric vehicle","authors":"Qian Mei ,&nbsp;Wenxia Guo ,&nbsp;Yanan Zhao ,&nbsp;Liming Nie ,&nbsp;Deepak Adhikari","doi":"10.1016/j.inffus.2024.102732","DOIUrl":"10.1016/j.inffus.2024.102732","url":null,"abstract":"<div><div>The emerging proportion of renewable energy resources penetration and the rapid popularity of Electric Vehicles (EVs) have promoted the development of the Internet of Electric Vehicles (IoEV), which enables seamless EV’ information collection and energy delivery by leveraging wireless power transfer. However, vulnerabilities in internet infrastructure and the self-interested behavior of EVs pose significant security and privacy risks during energy delivery in IoEV. In addition, EVs often lack the incentive to cooperate for regional energy balance. To tackle these questions, this paper proposes a blockchain-based privacy-preserving incentive mechanism for energy delivery in IoEV. Based on cryptographic technology, this paper introduces a group signature scheme with self-controlled and sequential linkability, which safeguards the privacy of EV users and ensures transaction records maintain exact sequence during energy delivery. Furthermore, an incentive mechanism based on co-utile reputation management is presented to encourage EV users to participate honestly and cooperatively in energy delivery. Moreover, a comprehensive security analysis of the proposed group signature scheme and incentive mechanism is given. Finally, extensive experimental results demonstrate the feasibility and efficiency of the proposed approach compared to existing schemes.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102732"},"PeriodicalIF":14.7,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142531895","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":"An efficient cross-view image fusion method based on selected state space and hashing for promoting urban perception","authors":"Peng Han ,&nbsp;Chao Chen","doi":"10.1016/j.inffus.2024.102737","DOIUrl":"10.1016/j.inffus.2024.102737","url":null,"abstract":"<div><div>In the field of cross-view image geolocation, traditional convolutional neural network (CNN)-based learning models generate unsatisfactory fusion performance due to their inability to model global correlations. The Transformer-based fusion methods can well compensate for the above problems, however, the Transformer has quadratic computational complexity and huge GPU memory consumption. The recent Mamba model based on the selection state space has a strong ability to model long sequences, lower GPU memory occupancy, and fewer GFLOPs. It is thus attractive and worth studying to apply Mamba to the cross-view image geolocation task. In addition, in the image-matching process (i.e., fusion of satellite/aerial and street view data.), we found that the storage occupancy of similarity measures based on floating-point features is high. Efficiently converting floating-point features into hash codes is a possible solution. In this study, we propose a cross-view image geolocation method (S6HG) based purely on Vision Mamba and hashing. S6HG fully utilizes the advantages of Vision Mamba in global information modeling and explicit location information encoding and the low storage occupancy of hash codes. Our method consists of two stages. In the first stage, we use a Siamese network based purely on vision Mamba to embed features for street view images and satellite images respectively. Our first-stage model is called S6G. In the second stage, we construct a cross-view autoencoder to further refine and compress the embedded features, and then simply map the refined features to hash codes. Comprehensive experiments show that S6G has achieved superior results on the CVACT dataset and comparable results to the most advanced methods on the CVUSA dataset. It is worth noting that other floating-point feature-based methods (4096-dimension) are 170.59 times faster than S6HG (768-bit) in storing 90,618 retrieval gallery data. Furthermore, the inference efficiency of S6G is higher than ViT-based computational methods.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102737"},"PeriodicalIF":14.7,"publicationDate":"2024-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142554837","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}