检索结果-内蒙古大学图书馆

Improving the Quality of Crowdsourced Image Labeling via LabelSimilarity

Journal of Computer Science & Technology 2017年第5期32卷 877-889页

作者： Yi-Li Fang Hai-Long Sun Peng-Peng Chen Ting Deng State Key Laboratory of Software Development Environment Beihang University Beijing 100191 China

Crowdsourcing is an effective method to obtain large databases of manually-labeled images, which is especially important for image understanding with supervised machine learning algorithms. However, for several kinds of tasks regarding image labeling, e.g., dog breed recognition, it is hard to achieve high-quality results. Therefore, further optimizing crowdsourcing workflow mainly involves task allocation and result inference. For task allocation, we design a two-round crowdsourcing framework, which contains a smart decision mechanism based on information entropy to determine whether to perform the second round task allocation. Regarding result inference, after quantifying the similarity of all labels, two graphical models are proposed to describe the labeling process and corresponding inference algorithms are designed to further improve the result quality of image labeling. Extensive experiments on real-world tasks in Crowdflower and synthesis datasets were conducted. The experimental results demonstrate the superiority of these methods in comparison with state-of-the-art methods.

关键词： image labeling crowdsourcing information entropy label similarity

来源：评论

学校读者我要写书评

暂无评论

Prune and Merge: Efficient Token Compression for Vision Transformer with Spatial Information Preserved

引用

IEEE Transactions on Multimedia 2025年

作者： Mao, Junzhu Shen, Yang Guo, Jinyang Yao, Yazhou Hua, Xiansheng Shen, Hengtao Nanjing University of Science and Technology School of Computer Science and Engineering Nanjing210094 China Beihang University State Key Laboratory of Software Development Environment Institute of Artificial Intelligence Beijing100191 China Terminus Group Beijing100027 China University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu611731 China

Token compression is essential for reducing the computational and memory requirements of transformer models, enabling their deployment in resource-constrained environments. In this work, we propose an efficient and hardware-compatible token compression method called Prune and Merge. Our approach integrates token pruning and merging operations within transformer models to achieve layer-wise token compression. By introducing trainable merge and reconstruct matrices and utilizing shortcut connections, we efficiently merge tokens while preserving important information and enabling the restoration of pruned tokens. Additionally, we introduce a novel gradient-weighted attention scoring mechanism that computes token importance scores during the training phase, eliminating the need for separate computations during inference and enhancing compression efficiency. We also leverage gradient information to capture the global impact of tokens and automatically identify optimal compression structures. Extensive experiments on the ImageNet-1k and ADE20K datasets validate the effectiveness of our approach, achieving significant speed-ups with minimal accuracy degradation compared to state-of-the-art methods. For instance, on DeiT-Small, we achieve a 1.64× speed-up with only a 0.2% drop in accuracy on ImageNet-1k. Moreover, by compressing segmenter models and comparing with existing methods, we demonstrate the superior performance of our approach in terms of efficiency and effectiveness. Code and models have been made available at https://***/NUST-Machine-Intelligence-laboratory/prune_and_merge. © 1999-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Dynamic evolution of collective emotions in social networks: a case study of Sina weibo

引用

Science China(Information Sciences) 2013年第7期56卷 150-167页

作者： XIONG XiaoBing ZHOU Gang HUANG YongZhong CHEN HaiYong XU Ke State Key Laboratory of Mathematical Engineering and Advanced Computing State Key Laboratory of Software Development Environment Beihang University

Emotion is a fundamental object of human existence and determined by a complex set of factors. With the rapid development of online social networks (OSNs), more and more people would like to express their emotion in OSNs, which provides wonderful opportunities to gain insight into how and why individual emotion is evolved in social network. In this paper, we focus on emotion dynamics in OSNs, and try to recognize the evolving process of collective emotions. As a basis of this research, we first construct a corpus and build an emotion classifier based on Bayes theory, and some effective strategies (entropy and salience) are introduced to improve the performance of our classifier, with which we can classify any Chinese tweet into a particular emotion with an accuracy as high as 82%. By analyzing the collective emotions in our sample networks in detail, we get some interesting findings, including a phenomenon of emotion synchronization between friends in OSNs, which offers good evidence for that human emotion can be spread from one person to another. Furthermore, we find that the number of friends has strong correlation with individual emotion. Based on those useful findings, we present a dynamic evolution model of collective emotions, in which both self-evolving process and mutual-evolving process are considered. To this end, extensive simulations on both real and artificial networks have been done to estimate the parameters of our emotion dynamic model, and we find that mutual-evolution plays a more important role than self-evolution in the distribution of collective emotions. As an application of our emotion dynamic model, we design an efficient strategy to control the collective emotions of the whole network by selecting seed users according to k-core rather than degree.

关键词： online social network emotion classifier emotion spread emotion dynamic model emotion controlling

来源：评论

学校读者我要写书评

暂无评论

Towards Label-Efficient Deep Learning-based Aging-related Bug Prediction with Spiking Convolutional Neural Networks

引用

IEEE Transactions on Emerging Topics in Computing 2025年

作者： Tian, Yunzhe Li, Yike Chen, Kang Zhang, Zhenguo Tong, Endong Liu, Jiqiang Qin, Fangyun Zheng, Zheng Niu, Wenjia Hebei Boshilin Technology Development Co. Ltd. Shijiazhuang050051 China Beijing Jiaotong University Beijing Key Laboratory of Security and Privacy in Intelligent Transportation Beijing100044 China Tangshan Research Institute of Beijing Jiaotong University Tangshan063000 China Capital Normal University College of Information Engineering Beijing100048 China Nanjing University State Key Laboratory for Novel Software Technology Nanjing210093 China Beihang University Laboratory of Dependable Intelligent Systems School of Automation Science and Electrical Engineering Beijing100191 China

Recent advances in Deep Learning (DL) have enhanced Aging-Related Bug (ARB) prediction for mitigating software aging. However, DL-based ARB prediction models face a dual challenge: overcoming overfitting to enhance generalization and managing the high labeling costs associated with extensive data requirements. To address the first issue, we utilize the sparse and binary nature of spiking communication in Spiking Neural Networks (SNNs), which inherently provides brain-inspired regularization to effectively alleviate overfitting. Therefore, we propose a Spiking Convolutional Neural Network (SCNN)-based ARB prediction model along with a training framework that handles the model's spatial-temporal dynamics and non-differentiable nature. To reduce labeling costs, we introduce a Bio-inspired and Diversity-aware Active Learning framework (BiDAL), which prioritizes highly informative and diverse samples, enabling more efficient usage of the limited labeling budget. This framework incorporates bio-inspired uncertainty to enhance informativeness measurement along with using a diversity-aware selection strategy based on clustering to prevent redundant labeling. Experiments on three ARB datasets show that ARB-SCNN effectively reduces overfitting, improving generalization performance by 6.65% over other DL-based classifiers. Additionally, BiDAL boosts label efficiency for ARB-SCNN training, outperforming four state-of-the-art active learning methods by 4.77% within limited labeling budgets. © 2013 IEEE.

关键词： Active learning

来源：评论

学校读者我要写书评

暂无评论

Paleolithic divergence and multiple Neolithic expansions of ancestral nomadic emperor-related paternal lineages

引用

Journal of Genetics and Genomics 2025年第4期52卷 502-512页

作者： Mengge Wang Qiuxia Sun Yuhang Feng Lan-Hai Wei Kaijun Liu Lintao Luo Yuguo Huang Kun Zhou Haibing Yuan Hongliang Lv Yu Lu Jing Cheng Shaoqing Wen Chuan-Chao Wang Renkuan Tang Fengxiao Bu Chao Liu Huijun Yuan Zhiyong Wang Guanglin He Institute of Rare Diseases Frontiers Science Center for Disease-related Molecular NetworkWest China HospitalSichuan UniversityChengduSichuan 610000China Center for Archaeological Science Sichuan UniversityChengduSichuan 610000China Anti-Drug Technology Center of Guangdong Province GuangzhouGuangdong 510230China Department of Forensic Medicine College of Basic MedicineChongqing Medical UniversityChongqing 400331China Human Genetics and Forensic Genomics Research Institute College of Basic MedicineChongqing Medical UniversityChongqing 400331China Institute of Humanities and Human Sciences Inner Mongolia Normal UniversityHohhotInner Mongolia 010022China Institute of 23Mofang Tianfu Software ParkChengduSichuan 610042China MOE Laboratory for National Development and Intelligent Governance Fudan UniversityShanghai 200433China State Key Laboratory of Cellular Stress Biology School of Life SciencesXiamen UniversityXiamenFujian 361102China

The reconstruction of demographic history using ancient and modern genomic resources reveals extensive interactions and admixture between ancient nomadic pastoralists and the social organizations of the Chinese Central ***,the extent to which Y-chromosome genetic legacies from nomadic emperor-related ancestral lineages influence the Chinese paternal gene pool remains ***,we genotype 2717 ethnolinguistically diverse samples belonging to C2a lineages,perform whole-genome sequencing on 997 representative samples,and integrate these data with ancient genomic *** reconstruct the evolutionary histories of Northern Zhou-,Qing emperor-,and pastoralist-related lineages to assess their genetic impact on modern Chinese *** reassembled fine-scale Ychromosome phylogeny identifies deep divergence and five Neolithic expansion events contributing differently to the formation of northern Chinese *** modeling indicates that the nomadic empires of the Northern Zhou and Qing dynasties genetically originated from the Mongolian *** topology and shared haplotype patterns show that three upstream ancestors of Northern Zhou(C2a1a1b1a2a1b-FGC28857),Donghu tribe(C2a1a1b1-F1756),and Qing(C2a1a3a2-F10283)emperor-related lineages expanded during the middle Neolithic,contributing significantly to genetic flow between ancient northeastern Asians and modern East ***,this study reveals limited direct contributions of Emperor Wu of Northern Zhou’s lineages to modern East Asians.

关键词： Paternal lineage Founding lineage Genetic origin Emperor lineage Y-chromosome

来源：评论

学校读者我要写书评

暂无评论

3D Anatomical Structure-Guided Deep Learning for Accurate Diffusion Microstructure Imaging

3D Anatomical Structure-Guided Deep Learning for Accurate Di...

引用

IEEE International Symposium on Biomedical Imaging

作者： Xinrui Ma Jian Cheng Wenxin Fan Ruoyou Wu Yongquan Ye Shanshan Wang Paul C. Lauterbur Research Center for Biomedical Imaging Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen Guangdong China University of Chinese Academy of Sciences Beijing China State Key Laboratory of Software Development Environment Beihang University Beijing China Key Laboratory of Data Science and Intelligent Computing Institute of International Innovation Beihang University Hangzhou Zhejiang China Pengcheng Laboratory Shenzhen Guangdong China United Imaging Healthcare Houston TX USA

ISBN: (数字)9798331520526

ISBN: (纸本)9798331520533

Diffusion magnetic resonance imaging (dMRI) is a crucial non-invasive technique for exploring the microstructure of the living human brain. Traditional hand-crafted and model-based tissue microstructure reconstruction methods often require extensive diffusion gradient sampling, which can be time-consuming and limits the clinical applicability of tissue microstructure information. Recent advances in deep learning have shown promise in microstructure estimation; however, accurately estimating tissue microstructure from clinically feasible dMRI scans remains challenging without appropriate constraints. This paper introduces a novel framework that achieves high-fidelity and rapid diffusion microstructure imaging by simultaneously leveraging anatomical information from macro-level priors and mutual information across parameters. This approach enhances time efficiency while maintaining accuracy in microstructure estimation. Experimental results demonstrate that our method outperforms four state-of-the-art techniques, achieving a peak signal-to-noise ratio (PSNR) of 30.51±0.58 and a structural similarity index measure (SSIM) of 0.97±0.004 in estimating parametric maps of multiple diffusion models. Notably, our method achieves a 15× acceleration compared to the dense sampling approach, which typically utilizes 270 diffusion gradients.

关键词： Deep learning PSNR Accuracy Three-dimensional displays Magnetic resonance imaging Estimation Reconstruction algorithms Diffusion models Feature extraction Microstructure

来源：评论

学校读者我要写书评

暂无评论

REPRESENTATION AND AUTOMATED TRANSFORMATION OF GEOMETRIC stateMENTS

引用

Journal of Systems Science & Complexity 2014年第2期27卷 382-412页

作者： CHEN Xiaoyu State Key Laboratory of Software Development Environment School of Computer Science and EngineeringBeihang University

The emergence of a large quantity of digital resources in geometry, various geometric automated theorem proving systems, and kinds of dynamic geometry software systems has made geometric computation, reasoning, drawing, and knowledge management dynamic, automatic or interactive on computer. Integration of electronic contents and different systems is desired to enhance their accessibility and exploitability. This paper proposes an equivalent transformation framework for manipulating geometric statements available in the literature by using geometry software systems. Such a framework works based on a newly designed geometry description language(GDL), in which geometric statements can be represented naturally and easily. The author discusses and presents key procedures of automatically transforming GDL statements into target system-native representations for *** author also demonstrates the framework by illustrating equivalent transformation processes and interfaces for compiling the transformation results into executable formats that can be interpreted by the target geometry software systems for automated theorem proving and dynamic diagram drawing.

关键词： Automated diagram drawing equivalent transformation formalized geometric statements geometric automated theorem proving.

来源：评论

学校读者我要写书评

暂无评论

Recommender systems based on ranking performance optimization

引用

Frontiers of Computer Science 2016年第2期10卷 270-280页

作者： Richong ZHANG Han BAO Hailong SUN Yanghao WANG Xudong LIU State Key Laboratory of Software Development Environment Beihang University B eijing 100191 China

The rapid development of online services and information overload has inspired the fast development of recommender systems, among which collaborative filtering algorithms and model-based recommendation approaches are wildly exploited. For instance, matrix factorization （MF） demonstrated successful achievements and advantages in assisting internet users in finding interested information. These existing models focus on the prediction of the users＇ ratings on unknown items. The performance is usually evaluated by the metric root mean square error （RMSE）. However, achieving good performance in terms of RMSE does not always guarantee a good ranking performance. Therefore, in this paper, we advocate to treat the recommendation as a ranking problem. Normalized discounted cumulative gain （NDCG） is chosen as the optimization target when evaluating the ranking accuracy. Specifically, we present three ranking-oriented recommender algorithms, NSME AdaMF and AdaNSME NSMF builds a NDCG approximated loss function for Matrix Factorization. AdaMF is based on an algorithm by adaptively combining component MF recommenders with boosting method. To combine the advantages of both algorithms, we propose AdaNSME which is a hybird of NSMF and AdaME and show the superiority in both ranking accuracy and model generalization. In addition, we compare our proposed approaches with the state-of-the-art recommendation algorithms. The comparison studies confirm the advantage of our proposed approaches.

关键词： recommender system matrix factorization,learning to rank

来源：评论

学校读者我要写书评

暂无评论

Accelerating MUS enumeration by inconsistency graph partitioning

引用

Science China(Information Sciences) 2019年第11期62卷 70-80页

作者： Jie LUO Shaofan LIU State Key Laboratory of Software Development Environment School of Computer Science and EngineeringBeihang University

The problem of finding minimal unsatisfiable subsets(MUSes) has been studied frequently because of its theoretical importance and wide range of applications in domains such as electronic design automation, software, and integrated circuit verification. In this paper, a method for accelerating the enumeration of MUSes based on inconsistency graph partitioning is proposed. First, an inconsistency graph of a set of clauses is constructed by extracting the inconsistency relations between literals of different clauses. In this paper, we show that by partitioning the inconsistency graph into small connected components through a vertex cut, the enumeration of MUSes in different components becomes independent and it is possible to compute them separately. Moreover, the MUSes of the original clause set can be constructed by merging the unit clauses in the MUSes of these connected components back into the clauses in the vertex cut. Experiments show that by integrating the acceleration method into the MARCO MUSes enumerator, there is a 2–3 times improvement in the average runtime of solved instances for randomly generated benchmarks. By integrating the acceleration method into itself as an MUS enumerator, there is another 3–4 times improvement when compared with the accelerated MARCO.

关键词： minimal unsatisfiable subsets inconsistency graph graph partition SAT UNSAT

来源：评论

学校读者我要写书评

暂无评论

Personalized Tag Recommendation Using Social Influence

引用

Journal of Computer Science & Technology 2012年第3期27卷 527-540页

作者：胡军王兵刘禹李德毅 State Key Laboratory of Software Development Environment Beihang University

Tag recommendation encourages users to add more tags in bridging the semantic gap between human concept and the features of media object, which provides a feasible solution for content-based multimedia information retrieval. In this paper, we study personalized tag recommendation in a popular online photo sharing site Flickr. Social relationship information of users is collected to generate an online social network. From the perspective of network topology, we propose node topological potential to characterize user＇s social influence. With this metric, we distinguish different social relations between users and find out those who really have influence on the target users. Tag recommendations are based on tagging history and the latent personalized preference learned from those who have most influence in user＇s social network. We evaluate our method on large scale real-world data. The experimental results demonstrate that our method can outperform the non-personalized global co-occurrence method and other two state-of-the-art personalized approaches using social networks. We also analyze the further usage of our approach for the cold-start problem of tag recommendation.

关键词： recommendation system social tagging personalization social network Flickr

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：