Molecular representation learning (MRL) is essential in certain applications including drug discovery and life science. Despite advancements in multiview and multimodal learning in MRL, existing models have explored only a limited range of perspectives, and the fusion of different views and modalities in MRL remains underexplored. Besides, obtaining the geometric conformer of molecules is not feasible in many tasks due to the high computational cost. Designing a general-purpose pertaining model for MRL is worthwhile yet challenging. This paper proposes a novel graph Transformer pretraining framework with fusion of node and graph views, along with the 2D topology and 3D geometry modalities of molecules, called MolGT. This MolGT model integrates node-level and graph-level pretext tasks on 2D topology and 3D geometry, leveraging a customized modality-shared graph Transformer that has versatility regarding parameter efficiency and knowledge sharing across modalities. Moreover, MolGT can produce implicit 3D geometry by leveraging contrastive learning between 2D topological and 3D geometric modalities. We provide extensive experiments and in-depth analyses, verifying that MolGT can (1) indeed leverage multiview and multimodal information to represent molecules accurately, and (2) infer nearly identical results using 2D molecules without requiring the expensive computation of generating conformers. Code is available on GitHub11https://github.com/robbenplus/MolGT..
{"title":"Pretraining graph transformer for molecular representation with fusion of multimodal information","authors":"Ruizhe Chen, Chunyan Li, Longyue Wang, Mingquan Liu, Shugao Chen, Jiahao Yang, Xiangxiang Zeng","doi":"10.1016/j.inffus.2024.102784","DOIUrl":"https://doi.org/10.1016/j.inffus.2024.102784","url":null,"abstract":"Molecular representation learning (MRL) is essential in certain applications including drug discovery and life science. Despite advancements in multiview and multimodal learning in MRL, existing models have explored only a limited range of perspectives, and the fusion of different views and modalities in MRL remains underexplored. Besides, obtaining the geometric conformer of molecules is not feasible in many tasks due to the high computational cost. Designing a general-purpose pertaining model for MRL is worthwhile yet challenging. This paper proposes a novel graph Transformer pretraining framework with fusion of node and graph views, along with the 2D topology and 3D geometry modalities of molecules, called MolGT. This MolGT model integrates node-level and graph-level pretext tasks on 2D topology and 3D geometry, leveraging a customized modality-shared graph Transformer that has versatility regarding parameter efficiency and knowledge sharing across modalities. Moreover, MolGT can produce implicit 3D geometry by leveraging contrastive learning between 2D topological and 3D geometric modalities. We provide extensive experiments and in-depth analyses, verifying that MolGT can (1) indeed leverage multiview and multimodal information to represent molecules accurately, and (2) infer nearly identical results using 2D molecules without requiring the expensive computation of generating conformers. Code is available on GitHub<ce:cross-ref ref><ce:sup loc=\"post\">1</ce:sup></ce:cross-ref><ce:footnote><ce:label>1</ce:label><ce:note-para view=\"all\"><ce:inter-ref xlink:href=\"https://github.com/robbenplus/MolGT\" xlink:type=\"simple\">https://github.com/robbenplus/MolGT</ce:inter-ref>.</ce:note-para></ce:footnote>.","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"36 1","pages":""},"PeriodicalIF":18.6,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670566","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}
Pub Date : 2024-11-08DOI: 10.1016/j.inffus.2024.102779
Xuanhua He , Ke Cao , Jie Zhang , Keyu Yan , Yingying Wang , Rui Li , Chengjun Xie , Danfeng Hong , Man Zhou
Pan-sharpening involves integrating information from low-resolution multi-spectral and high-resolution panchromatic images to generate high-resolution multi-spectral counterparts. While recent advancements in the state space model, particularly the efficient long-range dependency modeling achieved by Mamba, have revolutionized computer vision community, its untapped potential in pan-sharpening motivates our exploration. Our contribution, Pan-Mamba, represents a novel pan-sharpening network that leverages the efficiency of the Mamba model in global information modeling. In Pan-Mamba, we customize two core components: channel swapping Mamba and cross-modal Mamba, strategically designed for efficient cross-modal information exchange and fusion. The former initiates a lightweight cross-modal interaction through the exchange of partial panchromatic and multi-spectral channels, while the latter facilities the information representation capability by exploiting inherent cross-modal relationships. Through extensive experiments across diverse datasets, our proposed approach surpasses state-of-the-art methods, showcasing superior fusion results in pan-sharpening. To the best of our knowledge, this work is the first attempt in exploring the potential of the Mamba model and establishes a new frontier in the pan-sharpening techniques. The source code is available at https://github.com/alexhe101/Pan-Mamba.
{"title":"Pan-Mamba: Effective pan-sharpening with state space model","authors":"Xuanhua He , Ke Cao , Jie Zhang , Keyu Yan , Yingying Wang , Rui Li , Chengjun Xie , Danfeng Hong , Man Zhou","doi":"10.1016/j.inffus.2024.102779","DOIUrl":"10.1016/j.inffus.2024.102779","url":null,"abstract":"<div><div>Pan-sharpening involves integrating information from low-resolution multi-spectral and high-resolution panchromatic images to generate high-resolution multi-spectral counterparts. While recent advancements in the state space model, particularly the efficient long-range dependency modeling achieved by Mamba, have revolutionized computer vision community, its untapped potential in pan-sharpening motivates our exploration. Our contribution, Pan-Mamba, represents a novel pan-sharpening network that leverages the efficiency of the Mamba model in global information modeling. In Pan-Mamba, we customize two core components: channel swapping Mamba and cross-modal Mamba, strategically designed for efficient cross-modal information exchange and fusion. The former initiates a lightweight cross-modal interaction through the exchange of partial panchromatic and multi-spectral channels, while the latter facilities the information representation capability by exploiting inherent cross-modal relationships. Through extensive experiments across diverse datasets, our proposed approach surpasses state-of-the-art methods, showcasing superior fusion results in pan-sharpening. To the best of our knowledge, this work is the first attempt in exploring the potential of the Mamba model and establishes a new frontier in the pan-sharpening techniques. The source code is available at <span><span>https://github.com/alexhe101/Pan-Mamba</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102779"},"PeriodicalIF":14.7,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658115","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}
Pub Date : 2024-11-06DOI: 10.1016/j.inffus.2024.102756
Fahad Sabah , Yuwen Chen , Zhen Yang , Abdul Raheem , Muhammad Azam , Nadeem Ahmad , Raheem Sarwar
Personalized federated learning (PFL) addresses the significant challenge of non-independent and identically distributed (non-IID) data across clients in federated learning (FL). Our proposed framework, “FairDPFL-SCS: Fair Dynamic Personalized Federated Learning with Strategic Client Selection”, marks a notable advancement in this domain. By integrating dynamic learning rate adjustments and a strategic client selection mechanism, our approach effectively mitigates the challenges posed by non-IID data while enhancing model personalization, fairness, and efficiency. We evaluated FairDPFL-SCS using standard datasets, including MNIST, FashionMNIST, and SVHN, employing architectures like VGG and CNN. Our model achieved impressive results, attaining 99.04% accuracy on MNIST, 89.19% on FashionMNIST, and 90.9% on SVHN. These results represent a substantial improvement over existing methods, including a highest increase of 16.74% in accuracy on SVHN when compared to the best-performing benchmark methods. In particular, our method also demonstrated lower fairness variance, presenting the importance of fairness in model personalization, a frequently overlooked aspect in FL research. Through extensive experiments, we validate the superior performance of FairDPFL-SCS compared to benchmark PFL approaches, highlighting significant improvements over state-of-the-art methods. This work represents a promising step forward in the field of federated learning, offering a comprehensive solution to the challenges presented by non-IID data while prioritizing fairness and efficiency in model personalization.
{"title":"FairDPFL-SCS: Fair Dynamic Personalized Federated Learning with strategic client selection for improved accuracy and fairness","authors":"Fahad Sabah , Yuwen Chen , Zhen Yang , Abdul Raheem , Muhammad Azam , Nadeem Ahmad , Raheem Sarwar","doi":"10.1016/j.inffus.2024.102756","DOIUrl":"10.1016/j.inffus.2024.102756","url":null,"abstract":"<div><div>Personalized federated learning (PFL) addresses the significant challenge of non-independent and identically distributed (non-IID) data across clients in federated learning (FL). Our proposed framework, “FairDPFL-SCS: Fair Dynamic Personalized Federated Learning with Strategic Client Selection”, marks a notable advancement in this domain. By integrating dynamic learning rate adjustments and a strategic client selection mechanism, our approach effectively mitigates the challenges posed by non-IID data while enhancing model personalization, fairness, and efficiency. We evaluated FairDPFL-SCS using standard datasets, including MNIST, FashionMNIST, and SVHN, employing architectures like VGG and CNN. Our model achieved impressive results, attaining 99.04% accuracy on MNIST, 89.19% on FashionMNIST, and 90.9% on SVHN. These results represent a substantial improvement over existing methods, including a highest increase of 16.74% in accuracy on SVHN when compared to the best-performing benchmark methods. In particular, our method also demonstrated lower fairness variance, presenting the importance of fairness in model personalization, a frequently overlooked aspect in FL research. Through extensive experiments, we validate the superior performance of FairDPFL-SCS compared to benchmark PFL approaches, highlighting significant improvements over state-of-the-art methods. This work represents a promising step forward in the field of federated learning, offering a comprehensive solution to the challenges presented by non-IID data while prioritizing fairness and efficiency in model personalization.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102756"},"PeriodicalIF":14.7,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658118","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}
Timely and accurate diagnosis of bearing faults can effectively reduce the chance of accidents in equipment. However, deep learning methods are mostly completely dependent on data and lack interpretability. It is difficult to deal with the differences between real-time data and training data under changing working conditions and noisy environments. In this study, we proposed M-IPISincNet, an explainability multi-source physics-informed convolutional network based on improved SincNet. Rolling bearing fault diagnosis is realized by extracting fault features from vibration and current signals. Firstly, a physics-informed convolutional layer is designed based on inverse Fourier transform and bandpass filters. Fault features are extracted by multi-scale convolution and multi-layer nonlinear mapping. A DBN network is applied extract unsupervised hidden fusion features in the vibration and current signals. The proposed method is validated under the datasets of Paderborn University (PU) and Case Western Reserve University (CWRU), which proves that the proposed method has explainability, robustness and great accuracy under multiple working conditions and noises.
{"title":"M-IPISincNet: An explainable multi-source physics-informed neural network based on improved SincNet for rolling bearings fault diagnosis","authors":"Jingshu Zhong , Yu Zheng , Chengtao Ruan , Liang Chen , Xiangyu Bao , Lyu Lyu","doi":"10.1016/j.inffus.2024.102761","DOIUrl":"10.1016/j.inffus.2024.102761","url":null,"abstract":"<div><div>Timely and accurate diagnosis of bearing faults can effectively reduce the chance of accidents in equipment. However, deep learning methods are mostly completely dependent on data and lack interpretability. It is difficult to deal with the differences between real-time data and training data under changing working conditions and noisy environments. In this study, we proposed M-IPISincNet, an explainability multi-source physics-informed convolutional network based on improved SincNet. Rolling bearing fault diagnosis is realized by extracting fault features from vibration and current signals. Firstly, a physics-informed convolutional layer is designed based on inverse Fourier transform and bandpass filters. Fault features are extracted by multi-scale convolution and multi-layer nonlinear mapping. A DBN network is applied extract unsupervised hidden fusion features in the vibration and current signals. The proposed method is validated under the datasets of Paderborn University (PU) and Case Western Reserve University (CWRU), which proves that the proposed method has explainability, robustness and great accuracy under multiple working conditions and noises.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102761"},"PeriodicalIF":14.7,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658120","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}
Pub Date : 2024-11-06DOI: 10.1016/j.inffus.2024.102751
Nahid Hasan , Md. Golam Rabiul Alam , Shamim H. Ripon , Phuoc Hung Pham , Mohammad Mehedi Hassan
Concerns related to data privacy, security, and ethical considerations become more prominent as data volumes continue to grow. In contrast to centralized setups, where all data is accessible at a single location, model-based clustering approaches can be successfully employed in federated settings. However, this approach to clustering in federated settings is still relatively unexplored and requires further attention. As federated clustering deals with remote data and requires privacy and security to be maintained, it poses particular challenges as well as possibilities. While model-based clustering offers promise in federated environments, a robust model aggregation method is essential for clustering rather than the generic model aggregation method like Federated Averaging (FedAvg). In this research, we proposed an autoencoder-based clustering method by introducing a novel model aggregation method FednadamN, which is a fusion of Adam and Nadam optimization approaches in a federated learning setting. Therefore, the proposed FednadamN adopted the adaptive learning rates based on the first and second moments of gradients from Adam which offered fast convergence and robustness to noisy data. Furthermore, FednadamN also incorporated the Nesterov-accelerated gradients from Nadam to further enhance the convergence speed and stability. We have studied the performance of the proposed Autoencoder-based clustering methods on benchmark datasets and using the novel FednadamN model aggregation strategy. It shows remarkable performance gain in federated clustering in comparison to the state-of-the-art.
随着数据量的不断增长，与数据隐私、安全和道德考虑相关的问题变得越来越突出。与集中式设置相比，基于模型的聚类方法可以在联合设置中成功应用，因为在集中式设置中，所有数据都可以在单一位置访问。不过，这种在联合环境中进行聚类的方法相对来说仍未得到探索，需要进一步关注。由于联合聚类处理的是远程数据，并要求维护隐私和安全，因此它既带来了特殊的挑战，也带来了更多的可能性。虽然基于模型的聚类在联合环境中大有可为，但对于聚类来说，必须有一种稳健的模型聚合方法，而不是像联合平均（FedAvg）这样的通用模型聚合方法。在本研究中，我们提出了一种基于自动编码器的聚类方法，引入了一种新颖的模型聚合方法 FednadamN，它是联盟学习环境中 Adam 和 Nadam 优化方法的融合。因此，所提出的 FednadamN 采用了基于 Adam 梯度第一矩和第二矩的自适应学习率，从而提供了快速收敛性和对噪声数据的鲁棒性。此外，FednadamN 还采用了 Nadam 的内斯特罗夫加速梯度，进一步提高了收敛速度和稳定性。我们在基准数据集上研究了所提出的基于自动编码器的聚类方法的性能，并使用了新颖的 FednadamN 模型聚合策略。结果表明，与最先进的方法相比，联合聚类的性能有了明显提高。
{"title":"An autoencoder-based confederated clustering leveraging a robust model fusion strategy for federated unsupervised learning","authors":"Nahid Hasan , Md. Golam Rabiul Alam , Shamim H. Ripon , Phuoc Hung Pham , Mohammad Mehedi Hassan","doi":"10.1016/j.inffus.2024.102751","DOIUrl":"10.1016/j.inffus.2024.102751","url":null,"abstract":"<div><div>Concerns related to data privacy, security, and ethical considerations become more prominent as data volumes continue to grow. In contrast to centralized setups, where all data is accessible at a single location, model-based clustering approaches can be successfully employed in federated settings. However, this approach to clustering in federated settings is still relatively unexplored and requires further attention. As federated clustering deals with remote data and requires privacy and security to be maintained, it poses particular challenges as well as possibilities. While model-based clustering offers promise in federated environments, a robust model aggregation method is essential for clustering rather than the generic model aggregation method like Federated Averaging (FedAvg). In this research, we proposed an autoencoder-based clustering method by introducing a novel model aggregation method FednadamN, which is a fusion of Adam and Nadam optimization approaches in a federated learning setting. Therefore, the proposed FednadamN adopted the adaptive learning rates based on the first and second moments of gradients from Adam which offered fast convergence and robustness to noisy data. Furthermore, FednadamN also incorporated the Nesterov-accelerated gradients from Nadam to further enhance the convergence speed and stability. We have studied the performance of the proposed Autoencoder-based clustering methods on benchmark datasets and using the novel FednadamN model aggregation strategy. It shows remarkable performance gain in federated clustering in comparison to the state-of-the-art.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102751"},"PeriodicalIF":14.7,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658113","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}
Pub Date : 2024-11-05DOI: 10.1016/j.inffus.2024.102704
Quanbo Ge , Kai Lin , Zhongyuan Zhao
In a complex environment, a multi-sensor fusion algorithm can compensate for the limitations of a single sensor’s performance. In a distributed fusion algorithm, sensors need to transmit local estimates to a central coordinate system, and the existence of coordinate transformation uncertainty can undermine the performance of data transmission. Therefore, this paper proposes a multi-sensor distributed fusion method based on trustworthiness. Firstly, considering the presence of non-Gaussian conversion errors, a credibility-based multi-sensor fusion framework is constructed. Secondly, to address the difficulty in estimating conversion errors when measurement errors follow a non-Gaussian distribution, an optimization model is constructed based on actual measurement information to estimate the distribution of non-Gaussian conversion errors. Then, in response to the non-linear and non-Gaussian characteristics of the target optimization function, a particle swarm optimization algorithm based on trustworthiness adaptive weights is proposed to estimate the coordinate transformation errors. Finally, given the inconsistency in local estimates due to missing sensor measurements or significant errors in a non-Gaussian complex environment, a maximum correntropy consensus algorithm is proposed to avoid the trustworthiness calculation being affected by the current measurement errors, thereby improving the accuracy of the global estimation.
{"title":"Credibility-based multi-sensor fusion for non-Gaussian conversion error mitigation","authors":"Quanbo Ge , Kai Lin , Zhongyuan Zhao","doi":"10.1016/j.inffus.2024.102704","DOIUrl":"10.1016/j.inffus.2024.102704","url":null,"abstract":"<div><div>In a complex environment, a multi-sensor fusion algorithm can compensate for the limitations of a single sensor’s performance. In a distributed fusion algorithm, sensors need to transmit local estimates to a central coordinate system, and the existence of coordinate transformation uncertainty can undermine the performance of data transmission. Therefore, this paper proposes a multi-sensor distributed fusion method based on trustworthiness. Firstly, considering the presence of non-Gaussian conversion errors, a credibility-based multi-sensor fusion framework is constructed. Secondly, to address the difficulty in estimating conversion errors when measurement errors follow a non-Gaussian distribution, an optimization model is constructed based on actual measurement information to estimate the distribution of non-Gaussian conversion errors. Then, in response to the non-linear and non-Gaussian characteristics of the target optimization function, a particle swarm optimization algorithm based on trustworthiness adaptive weights is proposed to estimate the coordinate transformation errors. Finally, given the inconsistency in local estimates due to missing sensor measurements or significant errors in a non-Gaussian complex environment, a maximum correntropy consensus algorithm is proposed to avoid the trustworthiness calculation being affected by the current measurement errors, thereby improving the accuracy of the global estimation.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102704"},"PeriodicalIF":14.7,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658121","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}
Pub Date : 2024-11-04DOI: 10.1016/j.inffus.2024.102780
Jianhui Lv , Byung-Gyu Kim , B.D. Parameshachari , Adam Slowik , Keqin Li
In the era of big data and artificial intelligence, healthcare data fusion analysis has become difficult because of the large amounts and different types of sources involved. Traditional methods are ineffective at processing and examination procedures for such complex multi-sensors of hyperscale healthcare data. To address this issue, we propose a novel large model-driven approach for hyperscale healthcare data fusion analysis in complex multi-sensor multi-sensors. Our method integrates data from various medical sensors and sources, using large models to extract and fuse information from structured and unstructured healthcare data. Then, we integrate these features with structured data using a hierarchical residual connected LSTM network, enhancing the model's ability to capture local and global context. Furthermore, we introduce a dynamic ReLU activation function and attention mechanism that allow us to adjust the depth of our networks dynamically while focusing only on relevant information. The experiments on MIMIC-III and eICU-CRD datasets demonstrate the superiority of the proposed method in terms of accuracy, efficiency, and robustness compared to state-of-the-art methods. Therefore, the proposed method provides valuable insights into the potential of large model-driven approaches for tackling the challenges of hyperscale healthcare data fusion analysis in complex multi-sensors.
{"title":"Large model-driven hyperscale healthcare data fusion analysis in complex multi-sensors","authors":"Jianhui Lv , Byung-Gyu Kim , B.D. Parameshachari , Adam Slowik , Keqin Li","doi":"10.1016/j.inffus.2024.102780","DOIUrl":"10.1016/j.inffus.2024.102780","url":null,"abstract":"<div><div>In the era of big data and artificial intelligence, healthcare data fusion analysis has become difficult because of the large amounts and different types of sources involved. Traditional methods are ineffective at processing and examination procedures for such complex multi-sensors of hyperscale healthcare data. To address this issue, we propose a novel large model-driven approach for hyperscale healthcare data fusion analysis in complex multi-sensor multi-sensors. Our method integrates data from various medical sensors and sources, using large models to extract and fuse information from structured and unstructured healthcare data. Then, we integrate these features with structured data using a hierarchical residual connected LSTM network, enhancing the model's ability to capture local and global context. Furthermore, we introduce a dynamic ReLU activation function and attention mechanism that allow us to adjust the depth of our networks dynamically while focusing only on relevant information. The experiments on MIMIC-III and eICU-CRD datasets demonstrate the superiority of the proposed method in terms of accuracy, efficiency, and robustness compared to state-of-the-art methods. Therefore, the proposed method provides valuable insights into the potential of large model-driven approaches for tackling the challenges of hyperscale healthcare data fusion analysis in complex multi-sensors.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102780"},"PeriodicalIF":14.7,"publicationDate":"2024-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593605","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}
Pub Date : 2024-11-02DOI: 10.1016/j.inffus.2024.102775
Zhiming Liu , Jinhai Li , Xiao Zhang , Xizhao Wang
Missing multi-label learning is to address the problem of missing labels in multi-label datasets for multi-label classification tasks. Notably, the complex dependencies that typically exist between labels make accurate classification particularly challenging in the presence of missing labels. Some existing missing multi-label classification models often utilize feature selection to effectively recognize the dependencies between labels and features. However, they are ineffective at capturing hierarchical relationships of feature information, probably leading to a decline in prediction performance. To address this problem, this paper proposes a missing multi-label classification model based on multi-level stochastic concept clustering (MML-MSCC) to make dependencies between features and labels recognized more accurately and prediction performance better. In our model, optimal granularity selection is achieved through the global mutual information between features and labels, which makes the study of stochastic granule concept across multiple granularities. Furthermore, we utilize a stochastic concept clustering method to combine similar feature information for the purpose of making the missing label completion more reasonable. Note that stochastic granule concept clustering is performed with cross-granularity, thereby effectively capturing hierarchical relationships among feature information. Finally, to evaluate the performance of our model, we compare the MML-MSCC model with 9 existing missing multi-label classification models on 12 open datasets in terms of six evaluation metrics.
{"title":"Multi-level information fusion for missing multi-label learning based on stochastic concept clustering","authors":"Zhiming Liu , Jinhai Li , Xiao Zhang , Xizhao Wang","doi":"10.1016/j.inffus.2024.102775","DOIUrl":"10.1016/j.inffus.2024.102775","url":null,"abstract":"<div><div>Missing multi-label learning is to address the problem of missing labels in multi-label datasets for multi-label classification tasks. Notably, the complex dependencies that typically exist between labels make accurate classification particularly challenging in the presence of missing labels. Some existing missing multi-label classification models often utilize feature selection to effectively recognize the dependencies between labels and features. However, they are ineffective at capturing hierarchical relationships of feature information, probably leading to a decline in prediction performance. To address this problem, this paper proposes a missing multi-label classification model based on multi-level stochastic concept clustering (MML-MSCC) to make dependencies between features and labels recognized more accurately and prediction performance better. In our model, optimal granularity selection is achieved through the global mutual information between features and labels, which makes the study of stochastic granule concept across multiple granularities. Furthermore, we utilize a stochastic concept clustering method to combine similar feature information for the purpose of making the missing label completion more reasonable. Note that stochastic granule concept clustering is performed with cross-granularity, thereby effectively capturing hierarchical relationships among feature information. Finally, to evaluate the performance of our model, we compare the MML-MSCC model with 9 existing missing multi-label classification models on 12 open datasets in terms of six evaluation metrics.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102775"},"PeriodicalIF":14.7,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658116","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}
Pub Date : 2024-11-02DOI: 10.1016/j.inffus.2024.102776
Wei Guo , Hangjun Che , Man-Fai Leung , Long Jin , Shiping Wen
Incomplete multi-view clustering (IMVC) aims to address the clustering problem of multi-view data with partially missing samples and has received widespread attention in recent years. Most existing IMVC methods still have the following issues that require to be further addressed. They focus solely on the first-order correlation information among samples, neglecting the more intricate high-order connections. Additionally, these methods always overlook the noise or inaccuracies in the self-representation matrix. To address above issues, a novel method named Robust Mixed-order Graph Learning (RMoGL) is proposed for IMVC. Specifically, to enhance the robustness to noise, the self-representation matrices are separated into clean graphs and noise graphs. To capture complex high-order relationships among samples, the dynamic high-order similarity graphs are innovatively constructed from the recovered data. The clean graphs are endowed with mixed-order information and tend towards to obtain a consensus graph via a self-weighted manner. An efficient algorithm based on Alternating Direction Method of Multipliers (ADMM) is designed to solve the proposed RMoGL, and superior performance is demonstrated by compared with nine state-of-the-art methods across eight datasets. The source code of this work is available at https://github.com/guowei1314/RMoGL.
{"title":"Robust Mixed-order Graph Learning for incomplete multi-view clustering","authors":"Wei Guo , Hangjun Che , Man-Fai Leung , Long Jin , Shiping Wen","doi":"10.1016/j.inffus.2024.102776","DOIUrl":"10.1016/j.inffus.2024.102776","url":null,"abstract":"<div><div>Incomplete multi-view clustering (IMVC) aims to address the clustering problem of multi-view data with partially missing samples and has received widespread attention in recent years. Most existing IMVC methods still have the following issues that require to be further addressed. They focus solely on the first-order correlation information among samples, neglecting the more intricate high-order connections. Additionally, these methods always overlook the noise or inaccuracies in the self-representation matrix. To address above issues, a novel method named Robust Mixed-order Graph Learning (RMoGL) is proposed for IMVC. Specifically, to enhance the robustness to noise, the self-representation matrices are separated into clean graphs and noise graphs. To capture complex high-order relationships among samples, the dynamic high-order similarity graphs are innovatively constructed from the recovered data. The clean graphs are endowed with mixed-order information and tend towards to obtain a consensus graph via a self-weighted manner. An efficient algorithm based on Alternating Direction Method of Multipliers (ADMM) is designed to solve the proposed RMoGL, and superior performance is demonstrated by compared with nine state-of-the-art methods across eight datasets. The source code of this work is available at <span><span>https://github.com/guowei1314/RMoGL</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102776"},"PeriodicalIF":14.7,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142658229","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}
Pub Date : 2024-11-01DOI: 10.1016/j.inffus.2024.102771
Xinghua Liu , Xuan Shao , Ye Li
Shared Autonomous Mobility on Demand (SAMoD) is considered one of the most efficient modes of transportation for future cities and has thus gained significant attention. However, it may attract the ridership of public transportation (PT) systems, leading to negative externalities such as traffic congestion and environmental pollution. Greater social benefits can only be realized by seamlessly integrating SAMoD with PT systems, leveraging SAMoD’s flexibility and PT’s large-scale transport capacity. Therefore, this study considers the various complex potential interactions between SAMoD and PT (such as subways, BRT, and buses), including first and last-mile services and alternatives, and aims to investigate an optimization framework for network construction and passenger flow allocation in a SAMoD-PT integrated system to achieve an optimal balance between sustainability and efficiency. Specifically, we first applied a hierarchical weighted K-means clustering algorithm to cluster multi-source travel demands and used the Voronoi partition algorithm for regional division. Secondly, potential connections in the multi-modal transportation network were determined using a greedy triangulation algorithm. Subsequently, life cycle assessment and continuous approximation algorithms were employed to quantify environmental costs (including greenhouse gas emissions and energy consumption) as well as passenger and operator costs, respectively. Finally, we constructed a multi-objective optimization model and solved it using the weighted sum method, obtaining the Pareto frontier to balance sustainability and efficiency in the SAMoD-PT integrated system. The results show that the optimized SAMoD-PT integrated system can significantly reduce social costs, mitigate inter-modal competition effects, and ensure the central role of PT. This highlights the great potential of cooperation between SAMoD and PT. These findings provide valuable insights for developing countries on how to plan more efficient and environmentally friendly multi-modal urban transportation systems in the future.
{"title":"Eco-friendly integration of shared autonomous mobility on demand and public transit based on multi-source data","authors":"Xinghua Liu , Xuan Shao , Ye Li","doi":"10.1016/j.inffus.2024.102771","DOIUrl":"10.1016/j.inffus.2024.102771","url":null,"abstract":"<div><div>Shared Autonomous Mobility on Demand (SAMoD) is considered one of the most efficient modes of transportation for future cities and has thus gained significant attention. However, it may attract the ridership of public transportation (PT) systems, leading to negative externalities such as traffic congestion and environmental pollution. Greater social benefits can only be realized by seamlessly integrating SAMoD with PT systems, leveraging SAMoD’s flexibility and PT’s large-scale transport capacity. Therefore, this study considers the various complex potential interactions between SAMoD and PT (such as subways, BRT, and buses), including first and last-mile services and alternatives, and aims to investigate an optimization framework for network construction and passenger flow allocation in a SAMoD-PT integrated system to achieve an optimal balance between sustainability and efficiency. Specifically, we first applied a hierarchical weighted K-means clustering algorithm to cluster multi-source travel demands and used the Voronoi partition algorithm for regional division. Secondly, potential connections in the multi-modal transportation network were determined using a greedy triangulation algorithm. Subsequently, life cycle assessment and continuous approximation algorithms were employed to quantify environmental costs (including greenhouse gas emissions and energy consumption) as well as passenger and operator costs, respectively. Finally, we constructed a multi-objective optimization model and solved it using the weighted sum method, obtaining the Pareto frontier to balance sustainability and efficiency in the SAMoD-PT integrated system. The results show that the optimized SAMoD-PT integrated system can significantly reduce social costs, mitigate inter-modal competition effects, and ensure the central role of PT. This highlights the great potential of cooperation between SAMoD and PT. These findings provide valuable insights for developing countries on how to plan more efficient and environmentally friendly multi-modal urban transportation systems in the future.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"115 ","pages":"Article 102771"},"PeriodicalIF":14.7,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593604","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：