检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Liu, Jiduan Liu, Jiahao Wang, Qifan Wang, Jingang Wu, Wei Xian, Yunsen Zhao, Dongyan Chen, Kai Yan, Rui Wangxuan Institute of Computer Technology Peking University China Center for Data Science AAIS Peking University China Meituan China Meta AI United States National Key Laboratory of General Artificial Intelligence China BIGAI Beijing China School of Economics Peking University China Gaoling School of Artificial Intelligence Renmin University of China China Engineering Research Center of Next-Generation Intelligent Search and Recommendation Ministry of Education China

Unsupervised sentence representation learning is one of the fundamental problems in natural language processing with various downstream applications. Recently, contrastive learning has been widely adopted which derives high-quality sentence representations by pulling similar semantics closer and pushing dissimilar ones away. However, these methods fail to capture the fine-grained ranking information among the sentences, where each sentence is only treated as either positive or negative. In many real-world scenarios, one needs to distinguish and rank the sentences based on their similarities to a query sentence, e.g., very relevant, moderate relevant, less relevant, irrelevant, etc. In this paper, we propose a novel approach, RankCSE, for unsupervised sentence representation learning, which incorporates ranking consistency and ranking distillation with contrastive learning into a unified framework. In particular, we learn semantically discriminative sentence representations by simultaneously ensuring ranking consistency between two representations with different dropout masks, and distilling listwise ranking knowledge from the teacher. An extensive set of experiments are conducted on both semantic textual similarity (STS) and transfer (TR) tasks. Experimental results demonstrate the superior performance of our approach over several state-of-the-art baselines. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

A Method of Searching for Optimal Coalition Structure for Solving Resource Scheduling Problem of Overall Load Balancing in Edge computing Environments

引用

Journal of Physics: Conference Series 2020年第3期1550卷

作者： Hu Yanan Li Chunsheng Zhang Kejia School of Computer & Information Technology Northeast Petroleum University Daqing 163318 (PR China) Heilongjiang Provincial Key Laboratory of Oil Big data & Intelligent Analysis Daqing 163318 (PR China)

The cooperative computing resource scheduling problem in the distributed computing environment is usually transformed into the problem about how to find the optimal coalition structure. However, affected by factors such as high node density and large number of concurrent tasks in edge computing environments, it is difficult to formulate resource scheduling schemes for overall load balancing in a large strategic space by traditional methods. In order to solve this problem, we proposed a method of searching for optimal coalition structure based on discrete particle swarm optimization (DPSO). Firstly, we converted the resource scheduling problem of edge computing into an optimization problem model, constructed a new structure of coalition structure to express the resource scheduling scheme, and gave key elements such as optimization objectives, decision variables, and constraint conditions. Secondly, we improved DPSO and designed M-ary discrete particle swarm optimization (MDPSO): improved particle state determination method for adapting to the new expression of the coalition structure, and proposed a power series-based inertia weight computing method to improve the optimization speed and the optimal solution quality. Finally, we simulated real application scenarios and designed comparative experiments on the optimization speed, the optimal solution quality, stability, etc., which verifies that the method proposed in this paper has obvious advantages in solving the resource scheduling problem of overall load balancing in edge computing.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design of a high-performance full-system instrumentation

引用

Dongbei Daxue Xuebao/Journal of Northeastern University 2014年 35卷 129-134页

作者： Min, Qing-Hao Zhang, Zheng Wang, Hao-Jun Zhang, Wei-Hua Software School Fudan University Shanghai201203 China Shanghai Key Laboratory of Data Science Fudan University Shanghai201203 China Parallel Processing Institute Fudan University Shanghai201203 China State Key Laboratory of Mathematic Engineering and Advanced Computing PLA Information Engineering University Zhengzhou450001 China

Instrumentation is an important mechanism to collect run-time information of applications for further analysis like bottleneck analysis, hotspot analysis and even bug detection and revision. Currently there are lots of analyzers using instrumentation to collect their source data. However, most of these tools lack of full-system support and thus become inadequate for analyzing multi-core programs as they cannot provide full-system information like scheduling, system calls and inter-process communication, etc. Meanwhile, full-system instrumentation needs to collect OS and hardware information, resulting to performance degradation. In this paper, a high-performance full-system instrumentation method was designed for collecting full-system information as well as application information. The proposed method was implemented inside the QEMU full-system emulator, and 40 MIPS performance on average was achieved after optimization, which was appropriated for normal program analysis. ©, 2014, Northeastern University. All right reserved.

关键词： Full-system High-performance Instrumentation Multi-core

来源：评论

学校读者我要写书评

暂无评论

DULDA: Dual-domain Unsupervised Learned Descent Algorithm for PET image reconstruction

arXiv

引用

arXiv 2023年

作者： Hu, Rui Chen, Yunmen Kim, Kyungsang Rockenbach, Marcio Aloisio Bezerra Cavalcanti Li, Quanzheng Liu, Huafeng State Key Laboratory of Modern Optical Instrumentation Department of Optical Engineering Zhejiang University Hangzhou310027 China Department of Mathematics University of Florida GainesvilleFL32611 United States The Center for Advanced Medical Computing and Analysis Massachusetts General Hospital Harvard Medical School BostonMA02114 United States Center for Clinical Data Science Massachusetts General Brigham BostonMA02116 United States

Deep learning based PET image reconstruction methods have achieved promising results recently. However, most of these methods follow a supervised learning paradigm, which rely heavily on the availability of high-quality training labels. In particular, the long scanning time required and high radiation exposure associated with PET scans make obtaining these labels impractical. In this paper, we propose a dual-domain unsupervised PET image reconstruction method based on learned descent algorithm, which reconstructs high-quality PET images from sinograms without the need for image labels. Specifically, we unroll the proximal gradient method with a learnable l2,1 norm for PET image reconstruction problem. The training is unsupervised, using measurement domain loss based on deep image prior as well as image domain loss based on rotation equivariance property. The experimental results demonstrate the superior performance of proposed method compared with maximum-likelihood expectation–maximization (MLEM), total-variation regularized EM (EM-TV) and deep image prior based method (DIP). Copyright © 2023, The Authors. All rights reserved.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

MultiAdapt: A Neural Network Adaptation For Pruning Filters Base on Multi-layers Group

引用

Journal of Physics: Conference Series 2021年第1期1873卷

作者： Jie Yang Zhihong Xie Ping Li School of Computer and Communication Engineering Changsha University of Science and Technology Changsha Hunan Changsha 410114 China Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation Changsha University of Science and Technology Changsha Hunan Changsha 410114 China

Deep convolutional neural networks have been widely used in various AI applications. The most advanced neural networks are becoming deeper and wider, which has caused some large convolutional neural networks to exceed the size limit of the server or application. The pruning algorithm provides a way to reduce the size of the neural network while keeping the accuracy as high as possible. The automatic progressive pruning algorithm is one of the widely used pruning algorithms. The progressive pruning algorithm prunes a certain layer of the network in each iteration to reduce the sparsity while preserving the accuracy as much as possible. In this article, we design a new automatic progressive pruning algorithm named MultiAdapt. MultiAdapt combines the combination method and the greedy algorithm. This multi-layers progressive pruning method greatly increases the search space of the greedy algorithm, making it possible to obtain a better pruning network. We use MultiAdapt to prune large neural networks VGG-16 and ResNet. The experimental results show that the MultiAdapt algorithm is better than other mainstream methods in the balance of neural network model size and accuracy. For image classification tasks on the ImageNet dataset, our method achieved 88.72% and 90.55% TOP-5 accuracy on the 50% sparsity VGG-16 and ResNet, while obtaining nearly 2×reduction in parameters and floating point numbers. The operation is reduced, and the reduction is higher than the recent popular method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Paired Phase and Magnitude Reconstruction for Advanced Diffusion-Weighted Imaging

arXiv

引用

arXiv 2022年

作者： Qian, Chen Wang, Zi Zhang, Xinlin Shi, Boxuan Jiang, Boyu Tao, Ran Li, Jing Ge, Yuwei Kang, Taishan Lin, Jianzhong Guo, Di Qu, Xiaobo The Department of Electronic Science Biomedical Intelligent Cloud R&D Center Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance National Institute for Data Science in Health and Medicine Xiamen University China The United Imaging Healthcare Shanghai China Xingaoyi Medical Equipment Co. Ltd. Yuyao China Department of Radiology Zhongshan Hospital of Xiamen University School of Medicine Xiamen University Xiamen China The School of Computer and Information Engineering Xiamen University of Technology Xiamen China

Objective: Multi-shot interleaved echo planer imaging can obtain diffusion-weighted images (DWI) with high spatial resolution and low distortion, but suffers from ghost artifacts introduced by phase variations between shots. In this work, we aim at solving the challenging reconstructions under inter-shot motions between shots and a low signal-to-noise ratio. Methods: An explicit phase model with paired phase and magnitude priors is proposed to regularize the reconstruction (PAIR). The former prior is derived from the smoothness of the shot phase and enforced with low-rankness in the k-space domain. The latter explores similar edges among multi-b-value and multi-direction DWI with weighted total variation in the image domain. Results: Extensive simulation and in vivo results show that PAIR can remove ghost artifacts very well under a high number of shots (8 shots) and significantly suppress the noise under the ultra-high b-value (4000 s/mm2). Conclusion: The explicit phase model PAIR with complementary priors has a good performance on challenging reconstructions under inter-shot motions between shots and a low signal-to-noise ratio. Significance: PAIR has great potential in advanced clinical DWI applications and brain function research. Copyright © 2022, The Authors. All rights reserved.

关键词： Clinical research

来源：评论

学校读者我要写书评

暂无评论

HRTS: Hierarchical Rauch-Tung-Striebel Smoother With Online Learning Priors for EEG Denoising

Journal of Network Intelligence

引用

Journal of Network Intelligence 2024年第2期9卷 673-688页

作者： Wang, Chuan-Sheng Zhang, Ling Fu, Tian-Lin Chen, Zhao-Qi Zhang, Fu-Quan Department of Automatic Control Technical Polytechnic University of Catalonia Autonomous Region of Catalonia Barcelona Spain School of Computer and Data Science Minjiang University No.200 Xiyuangong Road Fuzhou University Town Fujian Province Fuzhou City China College of Computer and Big Data Fuzhou University No.2 Wulongjiang North Road Fuzhou University Town Fujian Province Fuzhou City China Digital Media Art Key Laboratory of Sichuan Province Sichuan Conservatory of Music Fuzhou Technology Innovation Center of Intelligent Manufacturing Information System Minjiang University No.200 Xiyuangong Road Fuzhou University Town Fujian Province Fuzhou City China Fujian Province University No.1 Campus New Village Longjiang Street Fujian Province Fuqing City China

Electroencephalogram (EEG) is a nonlinear signal that reflects the physio-logical state of the brain at different times, containing rich information. However, the possible interference during the collection and transmission process often leads to a large amount of unnecessary noise in the EEG signal. Traditional denoising methods may face difficulties in high-dimensional data processing and the inability to effectively handle different types of Gaussian white noise. To better eliminate the interference of Gaussian white noise, this article specifically adopts the Hierarchical Rauch-Tung-Striebel Smoother (HRTS) method. This method can effectively integrate the learned structural prior EEG signals into the state space model, thereby describing the relationship between EEG signals and noise. Capture the spatiotemporal characteristics of EEG signals through hierarchical modeling, and optimize the components of EEG signals through estimation and prediction of hidden variables. Finally, the mean square error (MSE) is used as an evaluation indicator to compare and evaluate the denoising results. Empirical research has shown that the HRTS algorithm can not only effectively reduce runtime and better process high-dimensional data, but also significantly improve the quality of EEG signals, effectively suppress noise interference, and more accurately reflect the characteristics of brain activity. Compared to denoising algorithms such as Kalman filtering and Kalman wavelet filtering, the HRTS algorithm has more advantages in denoising EEG signals. © 2024, J. Network Intell. All rigjts reserved.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

High Quality Global Routing Algorithm Based on Multi-Stage Optimization

引用

Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics 2024年第4期36卷 607-614页

作者： Genggeng, Liu Zhenyu, Pei Ning, Xu College of Computer and Data Science Fuzhou University Fuzhou350116 China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China School of Information Engineering Wuhan University of Technology Wuhan430070 China

Global routing is a crucial step in VLSI physical design. To address the problems that the low utilization rate of capacity and easy to fall into local optimum in the existing global routing algorithms, a high-quality global routing algorithm based on multi-stage optimization is proposed. First, the virtual capacity dynamic adjustment strategy is introduced in the pre-connection stage and the routing reorganization stage to improve the utilization rate of capacity. Then, in the routing reorganization stage, the routing subarea adaptive expansion strategy is proposed to narrow the search space and improve the routing efficiency. Finally, the A* algorithm heuristic search is combined with the global search capability of the particle swarm algorithm to avoid the algorithm to choose the shortest path in the routing process, and thus avoid the algorithm to fall into the local optimum. Experimental results comparing with existing global routing algorithms on ISPD benchmark circuits show that the routing results of proposed algorithm reduce the wirelength by 6.0% to 22.8%. © 2024 Institute of computing Technology. All rights reserved.

关键词： Routing algorithms

来源：评论

学校读者我要写书评

暂无评论

Transformers in Medical Image Analysis: A Review

arXiv

引用

arXiv 2022年

作者： He, Kelei Gan, Chen Li, Zhuoyuan Rekik, Islem Yin, Zihao Ji, Wen Gao, Yang Wang, Qian Zhang, Junfeng Shen, Dinggang Medical School of Nanjing University Nanjing China National Institute of Healthcare Data Science Nanjing University Nanjing China State Key Laboratory for Novel Software Technology Nanjing University Nanjing China BASIRA Lab Faculty of Computer and Informatics Engineering Istanbul Technical University Istanbul Turkey School of Science and Engineering Computing University of Dundee United Kingdom School of Biomedical Engineering ShanghaiTech University Shanghai China Department of Research and Development Shanghai United Imaging Intelligence Co. Ltd. Shanghai China

Transformers have dominated the field of natural language processing, and recently impacted the computer vision area. In the field of medical image analysis, Transformers have also been successfully applied to full-stack clinical applications, including image synthesis/reconstruction, registration, segmentation, detection, and diagnosis. Our paper aims to promote awareness and application of Transformers in the field of medical image analysis. Specifically, we first overview the core concepts of the attention mechanism built into Transformers and other basic components. Second, we review various Transformer architectures tailored for medical image applications and discuss their limitations. Within this review, we investigate key challenges revolving around the use of Transformers in different learning paradigms, improving the model efficiency, and their coupling with other techniques. We hope this review can give a comprehensive picture of Transformers to the readers in the field of medical image analysis. © 2022, CC BY-SA.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Hierarchical deep CNN feature set-based representation learning for robust cross-resolution face recognition

arXiv

引用

arXiv 2021年

作者： Gao, Guangwei Yu, Yi Yang, Jian Qi, Guo-Jun Yang, Meng Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing210023 China Digital Content and Media Sciences Research Division National Institute of Informatics Tokyo101-8430 Japan School of Computer Science and Technology Nanjing University of Science and Technology Nanjing210094 China Department of Computer Science University of Central Florida OrlandoFL32816 United States School of Data and Computer Science Sun Yat-sen University Guangzhou510006 China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Sun Yat-sen University Guangzhou510006 China

Cross-resolution face recognition (CRFR), which is important in intelligent surveillance and biometric forensics, refers to the problem of matching a low-resolution (LR) probe face image against high-resolution (HR) gallery face images. Existing shallow learning-based and deep learning-based methods focus on mapping the HR-LR face pairs into a joint feature space where the resolution discrepancy is mitigated. However, little works consider how to extract and utilize the intermediate discriminative features from the noisy LR query faces to further mitigate the resolution discrepancy due to the resolution limitations. In this study, we desire to fully exploit the multilevel deep convolutional neural network (CNN) feature set for robust CRFR. In particular, our contributions are threefold. (i) To learn more robust and discriminative features, we desire to adaptively fuse the contextual features from different layers. (ii) To fully exploit these contextual features, we design a feature set-based representation learning (FSRL) scheme to collaboratively represent the hierarchical features for more accurate recognition. Moreover, FSRL utilizes the primitive form of feature maps to keep the latent structural information, especially in noisy cases. (iii) To further promote the recognition performance, we desire to fuse the hierarchical recognition outputs from different stages. Meanwhile, the discriminability from different scales can also be fully integrated. By exploiting these advantages, the efficiency of the proposed method can be delivered. Experimental results on several face datasets have verified the superiority of the presented algorithm to the other competitive CRFR approaches. Copyright © 2021, The Authors. All rights reserved.

关键词： Face recognition

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：