Information Fusion最新文献

{"title":"Knowledge fusion in deep learning-based medical vision-language models: A review","authors":"Dexuan Xu ,&nbsp;Yanyuan Chen ,&nbsp;Zhongyan Chai ,&nbsp;Yifan Xiao ,&nbsp;Yandong Yan ,&nbsp;Weiping Ding ,&nbsp;Hanpin Wang ,&nbsp;Zhi Jin ,&nbsp;Wenpin Jiao ,&nbsp;Weihua Yue ,&nbsp;Hang Li ,&nbsp;Yu Huang","doi":"10.1016/j.inffus.2025.103455","DOIUrl":"10.1016/j.inffus.2025.103455","url":null,"abstract":"<div><div>Medical vision-language models based on deep learning can automatically extract image features and fuse them with text information, which has promoted the rapid development of multimodal medical artificial intelligence. However, the complexity of the medical field requires the model to have a deep professional knowledge background. Therefore, knowledge fusion technology provides a new idea for solving medical vision-language tasks. Different from the existing reviews, this paper systematically sorts out the knowledge fusion methods in medical vision-language models from two unique perspectives: the stage characteristics of knowledge fusion and the task-oriented fusion strategy, and provides a new theoretical framework for research in the field. Firstly, this paper introduces the classification of medical knowledge and its applicable scenarios in detail. Subsequently, we systematically discuss the knowledge fusion algorithm based on deep learning and summarize the four different knowledge fusion stages (data construction, pretraining, feature representation and inference) in the medical vision-language model. In addition, this paper comprehensively analyzes the specific strategies of knowledge fusion in five types of medical vision-language tasks (medical report generation, medical visual question answering, medical language-guided segmentation, medical multimodal pretraining, and multimodal large language model), and summarizes the evaluation methods based on knowledge fusion in detail. Finally, we summarize future research directions, including enhanced interpretability, mixture-of-experts models, knowledge editing, etc., aiming to provide researchers with references that have both theoretical value and practical significance.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103455"},"PeriodicalIF":14.7,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579925","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":"GLIR-GaitNet: Global-Local Interaction Rebalancing Network for gait-based disease diagnosis","authors":"Yuanyuan Sun ,&nbsp;Qing Yang ,&nbsp;Xinyu Ji ,&nbsp;Yuyan Zhang ,&nbsp;Siyi Yu ,&nbsp;Meng Si ,&nbsp;Yuanyuan Xiang ,&nbsp;Bing Ji","doi":"10.1016/j.inffus.2025.103456","DOIUrl":"10.1016/j.inffus.2025.103456","url":null,"abstract":"<div><div>Gait analysis offers a promising approach for disease diagnosis, as various neurological and musculoskeletal disorders often lead to distinct gait patterns that can be systematically analyzed to assist in accurate clinical decision-making. While deep learning methods have been widely applied to multivariate gait time series, challenges remain—particularly in extracting complex coupling relationships across and within joints, as well as mitigating interference caused by asynchronous feature extraction during multi-feature fusion.</div><div>In this paper, the Global-Local Interaction Rebalancing Gait Network (GLIR-GaitNet) is proposed to solve these challenges, incorporating two novel modules. The Global-Local Joint Coupling Feature Extractor (GL-JCFE) module integrates a residual-based 2D local representation with a dynamic graph-based global modeling approach to capture comprehensive inter- and intra-joint coupling relationships. Within this module, the Multi-feature Cross Enhancement (MCE) is further introduced to strengthen feature complementarity across domains. Additionally, the Prototypical Interaction Rebalance (PIR) module enhances the consistency of feature distribution by introducing cross-domain similarity loss monitoring while reducing interference between feature extraction processes.</div><div>Extensive experiments conducted on three self-collected datasets (CSM-PS, LDH-LSS and CSM-HC) and one publicly available dataset (HOA) demonstrate that GLIR-GaitNet significantly outperforms state-of-the-art classification methods in terms of diagnostic average ACC and AUC. These results highlight the strong capability of the GLIR-GaitNet in distinguishing easily confused gait-related diseases and fine-grained severity grading. Our Code is available at this repository: <span><span>https://github.com/ginasmithe/GLIR-GaitNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103456"},"PeriodicalIF":14.7,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144588120","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":"Not all tokens are created equal: Perplexity Attention Weighted Networks for AI-generated text detection","authors":"Pablo Miralles-González,&nbsp;Javier Huertas-Tato,&nbsp;Alejandro Martín,&nbsp;David Camacho","doi":"10.1016/j.inffus.2025.103465","DOIUrl":"10.1016/j.inffus.2025.103465","url":null,"abstract":"<div><div>The rapid advancement in large language models (LLMs) has enhanced their ability to generate coherent text, raising concerns about misuse and making detection critical. However, this task remains challenging, particularly in unseen domains or with unfamiliar LLMs. Leveraging next-token distribution outputs offers a theoretically appealing detection approach, as they reflect the models’ extensive pre-training. Despite this promise, zero-shot methods using these outputs have had limited success. We hypothesize this is partly due to using the mean to aggregate next-token metrics across tokens, ignoring that some tokens are inherently easier or harder to predict. We propose the Perplexity Attention Weighted Network (PAWN), which uses last hidden states and positions to weight features derived from next-token distribution metrics across the sequence. Though not zero-shot, our method allows caching hidden states and metrics to reduce training resource needs. PAWN performs competitively or better in-distribution than the strongest baselines (fine-tuned LMs) using a fraction of their trainable parameters. It also generalizes better to unseen domains and source models, with less variability in the decision boundary under distribution shifts. PAWN is more robust to adversarial attacks and, with a multilingual backbone, generalizes well to languages unseen in supervised training, with LLaMA3-1B reaching a mean macro F1 score of 81.46% in cross-validation across nine languages.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103465"},"PeriodicalIF":14.7,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144588076","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":"Anatomic-kinetic fusion network for pituitary gland and microadenoma segmentation in multi-phase DCE-MRI","authors":"Xiangyu Li ,&nbsp;Yifan Liu ,&nbsp;Manxi Xu ,&nbsp;Hongwei Yu ,&nbsp;Jixin Luan ,&nbsp;Aocai Yang ,&nbsp;Bing Liu ,&nbsp;Amir Shmuel ,&nbsp;Yuan Zhen ,&nbsp;Gongning Luo ,&nbsp;Wei Wang ,&nbsp;Guolin Ma ,&nbsp;Kuanquan Wang","doi":"10.1016/j.inffus.2025.103457","DOIUrl":"10.1016/j.inffus.2025.103457","url":null,"abstract":"<div><div>Accurate segmentation of the pituitary gland and pituitary microadenoma (PM) using dynamic contrast-enhanced MRI (DCE-MRI) is crucial in clinical practice. However, effectively leveraging multi-phase DCE-MRI data for precise segmentation remains challenging due to inter-phase variability, missing anatomic details, and the disparities between anatomic and kinetic features in DCE-MRI. In this work, we propose a novel Anatomic-Kinetic Fusion Network (AKF-Net) to address these problems. The proposed AKF-Net makes four main contributions that have been proved to be quite effective in various experiments. First, we introduce a dynamic recursive graph learning method that enables kinetic dependencies modeling between phases while preserving anatomic consistency by iteratively refining its understanding of the relationships between voxels. Second, we propose an adaptive static-dynamic blending module that effectively improves the anatomic feature modeling by integrating multi-phase kinetic information with anatomic features in a residual attention manner. Third, we propose a novel anatomic-kinetic attention fusion module, which resolves the disparities between anatomic and kinetic features by effectively aligning and integrating spatial–temporal features in DCE-MRI, resulting in a robust fused representation that balances spatial detail and temporal dynamics. Fourth, to the best of our knowledge, the proposed AKF-Net is the first work to effectively model hemodynamics features in DCE-MRI by applying spatial–temporal graph learning based on the differential data. The experimental results demonstrate that AKF-Net outperforms existing segmentation methods, achieving state-of-the-art performance. The code is available at <span><span>https://github.com/PerceptionComputingLab/AKF-Net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103457"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144588121","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-Scale Data Fusion and AdaptiveLoss Kolmogorov–Arnold Network for multivariate time series forecasting","authors":"Jian Liu ,&nbsp;Fan Yang ,&nbsp;Ke Yan","doi":"10.1016/j.inffus.2025.103454","DOIUrl":"10.1016/j.inffus.2025.103454","url":null,"abstract":"<div><div>Real-world multivariate time series often exhibit multiple interwoven and highly coupled periodic patterns, along with significant volatility and uncertainty, which present substantial challenges for accurate time series forecasting. Inspired by the concept of multi-scale data fusion and the Kolmogorov–Arnold theory, this study proposes a novel time series forecasting framework that achieves both high predictive accuracy and parameter efficiency. The proposed approach comprises two key modules: the Multi-Scale Data Fusion Model (MDFM) and the AdaptiveLoss Kolmogorov–Arnold Network (AdaKAN). MDFM implements a two-stage fusion process that prioritizes trend and seasonal information across different temporal scales, thereby enhancing the model’s ability to capture both long-term trends and fine-grained fluctuations. AdaKAN leverages Gaussian radial basis functions and adaptive loss functions, significantly improving both the computational efficiency and predictive performance of the algorithm compared to the traditional Kolmogorov–Arnold Network. The experimental results indicate that, compared with recently published cutting-edge methods, MDFM-AdaKAN consistently demonstrated superior accuracy and adaptability. Specifically, it achieved a 15.2% improvement in the solar irradiance prediction task and a 19.4% improvement in the electricity transformer temperature prediction task, providing a reliable solution for high-precision time-series predictions.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103454"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144562886","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":"Choir-IDS: A federated learning framework for fidelity-calibrated explainable intrusion detection system for edge-IoT networks","authors":"Jyoti Prakash Sahoo ,&nbsp;Binayak Kar ,&nbsp;Ahmed M. Abdelmoniem ,&nbsp;Dimitris Chatzopoulos","doi":"10.1016/j.inffus.2025.103473","DOIUrl":"10.1016/j.inffus.2025.103473","url":null,"abstract":"<div><div>The proliferation of the Internet of Things (IoT) and its application in various domains has increased the need to detect malicious traffic in IoT networks. Machine learning-based intrusion detection systems (ML-IDS) have shown potential in addressing this challenge. However, ensuring the fidelity and explainability that contribute to the trustworthiness of ML-based IDS is essential for their effective deployment in real-world scenarios. Federated learning emerges as a promising approach to enhance privacy and scalability in ML-IDS by enabling collaborative model training across decentralized data sources. Ensemble learning techniques, when combined with federated learning, can further improve the performance of ML-based IDS. The objective of this study is to assess the performance of ensemble learning-based intrusion detection systems (IDS) in terms of fidelity and explainability within a federated learning framework. This research contributes to developing cutting-edge machine learning-based IDS techniques and offers valuable insights into the practical application of ensemble learning-based IDS in IoT and edge networks. With a specific emphasis on fidelity and explainability, this study introduces Choir-IDS, an advanced IDS framework specifically designed for edge-IoT environments. Choir-IDS draws inspiration from the notion of a “choir,” epitomizing harmonious collaboration. By harnessing the collaborative synergy of multiple machine learning models through federated learning, we introduce “Neural Boosting Ensembles,” which incorporate our approach to augmenting intrusion detection performance using deep learning and ensemble learning. By integrating advanced calibration techniques, including isotonic and sigmoid calibration, Choir-IDS ensures precise probability estimates tailored for diverse operational requirements. Furthermore, it incorporates cutting-edge model explainability tools, such as Local Interpretable Model-agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP), to enhance transparency and trust. This research establishes a new benchmark for fidelity-ensured and explainable IDS, offering practical insights into the application of federated learning in securing IoT and edge networks.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103473"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572765","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":"Structure-guided deep multi-view clustering","authors":"Jinrong Cui ,&nbsp;Xiaohuang Wu ,&nbsp;Haitao Zhang ,&nbsp;Chongjie Dong ,&nbsp;Jie Wen","doi":"10.1016/j.inffus.2025.103461","DOIUrl":"10.1016/j.inffus.2025.103461","url":null,"abstract":"<div><div>Deep multi-view clustering seeks to utilize the abundant information from multiple views to improve clustering performance. However, most of the existing clustering methods often neglect to fully mine multi-view structural information and fail to explore the distribution of multi-view data, limiting clustering performance. To address these limitations, we propose a structure-guided deep multi-view clustering model. Specifically, we introduce a positive sample selection strategy based on neighborhood relationships, coupled with a corresponding loss function. This strategy constructs multi-view nearest neighbor graphs to dynamically redefine positive sample pairs, enabling the mining of local structural information inside multi-view data and enhancing the reliability of positive sample selection. Additionally, we introduce a Gaussian distribution model to uncover latent structural information and introduce a loss function to reduce discrepancies between view embeddings. These two strategies explore multi-view structural information and data distribution from different perspectives, enhancing consistency across views and increasing intra-cluster compactness. Experimental evaluations demonstrate the efficacy of our method, showing significant improvements in clustering performance on multiple benchmark datasets compared to state-of-the-art multi-view clustering approaches.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103461"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144569989","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":"MVFusion-TSC: A multi-view fusion image-based network for time series classification","authors":"Chao Lian ,&nbsp;Yafeng Kang ,&nbsp;Wenjing Li ,&nbsp;Dongyu Zhou ,&nbsp;Tianang Sun ,&nbsp;Xiaoyong Lyu ,&nbsp;Yuliang Zhao","doi":"10.1016/j.inffus.2025.103458","DOIUrl":"10.1016/j.inffus.2025.103458","url":null,"abstract":"<div><div>Time series classification (TSC) is a supervised task that aims to categorize time series data into predefined classes by identifying meaningful patterns within their temporal structure. Current methods face numerous challenges in handling time-dependent modeling, capturing long- and short-term relationships, and feature extraction. To address these issues, we propose a time series classification framework, MVFusion-TSC, based on multi-view fusion image representation (MVFI) and a dual-branch feature fusion network (DFFNet). This framework uses the MVFI method to simultaneously encode local texture variations and global structural trends, and uses DFFNet, where one branch focuses on fine-grained local patterns while the other captures holistic temporal structures, effectively revealing and extracting both short- and long-term temporal dependencies. First, MVFI method is employed to convert time series data into multi-view fusion (MVF) images that integrate structural, color, and texture information. This enables the time-varying amplitude features of time-series data to be represented as structural and texture patterns in images, which better capture temporal dependencies and improve the performance of deep learning models. Secondly, DFFNet network is constructed, including a local feature extraction branch, a global feature extraction branch, and a gating fusion module, to achieve the extraction, fusion, and classification of both global and local temporal dependency features. Experiments on 25 benchmark time-series datasets show that the proposed method outperforms the current state-of-the-art methods in key metrics such as accuracy, win rate, and ranking, verifying its effectiveness and advantages in time-series classification tasks. The proposed method can provide a new research perspective for time-series classification, with significant theoretical value and practical application potential.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103458"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570039","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":"HIDFlowNet: A flow-based deep network for hyperspectral image denoising","authors":"Qizhou Wang ,&nbsp;Li Pang ,&nbsp;Xiangyong Cao ,&nbsp;Zhiqiang Tian ,&nbsp;Deyu Meng","doi":"10.1016/j.inffus.2025.103459","DOIUrl":"10.1016/j.inffus.2025.103459","url":null,"abstract":"<div><div>Hyperspectral image (HSI) denoising is essentially ill-posed since a noisy HSI can be degraded from multiple clean HSIs. However, existing deep learning (DL)-based approaches only restore one clean HSI from the given noisy HSI with a deterministic mapping, thus ignoring the ill-posed issue and always resulting in an over-smoothing problem. Additionally, these DL-based methods often neglect that noise is part of the high-frequency component and their network architectures fail to decouple the learning of low-frequency and high-frequency. To alleviate these issues, this paper proposes a flow-based HSI denoising network (HIDFlowNet) to directly learn the conditional distribution of the clean HSI given the noisy HSI and thus diverse clean HSIs can be sampled from the conditional distribution. Overall, our HIDFlowNet is induced from the generative flow model and is comprised of an invertible decoder and a conditional encoder, which can explicitly decouple the learning of low-frequency and high-frequency information of HSI. Specifically, the invertible decoder is built by stacking a succession of invertible conditional blocks (ICBs) to capture the local high-frequency details. The conditional encoder utilizes down-sampling operations to obtain low-resolution images and uses transformers to capture correlations over a long distance so that global low-frequency information can be effectively extracted. Extensive experiments on simulated and real HSI datasets verify that our proposed HIDFlowNet can obtain better or comparable results compared with other state-of-the-art methods.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103459"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144614043","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-task prediction of wind speed and time-varying wind shear coefficient using dynamic graph interactive neural network","authors":"Ke Fu,&nbsp;Zhengru Ren","doi":"10.1016/j.inffus.2025.103478","DOIUrl":"10.1016/j.inffus.2025.103478","url":null,"abstract":"<div><div>In the field of wind energy utilization, low-altitude wind speed and wind shear coefficient serve as pivotal variables for wind speed extrapolation, and thus, the wind speed at hub height or specific height can be effectively inferred. Although the power law model is widely used to describe wind profiles, traditional studies often assume that the wind shear coefficient is constant, typically 1/7. This simplification ignores the dynamic changes of the wind shear coefficient and potentially lead to prediction errors. To solve this problem, this study proposed an innovative multi-task prediction method using dynamic graph interactive neural network(DGINet), and the proposed method supports parallel computing. The novelty of this study lies in fully considering the time-varying characteristics of the wind shear coefficient and can accurately predict the wind speed and wind shear coefficient at the same time, so as to more accurately construct a vertical wind profile. The proposed DGINet consists of idealized sub-networks simplified to individual neuron and backbone network adopting the encoder–decoder architecture. The proposed encoder includes modified sample reconstruction strategy within the sliding window, which expands the data dimension, and fuses the improved gated graph unit with the cross convolution operator to model and perceive the multi-level correlation between samples. The experimental results show that the proposed model can accurately and simultaneously predict wind speed and wind shear coefficient within the prediction horizons of 15 min, 30 min and 1 h.</div></div>","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"125 ","pages":"Article 103478"},"PeriodicalIF":14.7,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572761","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}