检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Liu, Haixiong Li, Zuoyong Wu, Jiawei Zeng, Kun Hu, Rong Zeng, Wei Fujian Provincial Key Laboratory of Big Data Mining and Applications School of Computer Science and Mathematics Fujian University of Technology Fuzhou350118 China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control College of Computer and Data Science Minjiang University Fuzhou350121 China School of Intelligent Systems Engineering Shenzhen Campus of Sun Yat-sen University Guangdong Shenzhen518107 China The Key Laboratory of Cognitive Computing and Intelligent Information Processing of Fujian Education Institutions Wuyi University Wuyishan354300 China School of Physics and Mechanical and Electrical Engineering Longyan University Longyan364012 China

SSL(Semi-supervised learning) is widely used in machine learning, which leverages labeled and unlabeled data to improve model performance. SSL aims to optimize class mutual information, but noisy pseudo-labels introduce false class information due to the scarcity of labels. Therefore, these algorithms often need significant training time to refine pseudo-labels for performance improvement iteratively. To tackle this challenge, we propose a novel plug-and-play method named ASCL(Accelerating Semi-Supervised Learning via Contrastive Learning). This method combines contrastive learning with uncertainty-based selection for performance improvement and accelerates the convergence of SSL algorithms. Contrastive learning initially emphasizes the mutual information between samples as a means to decrease dependence on pseudo-labels. Subsequently, it gradually turns to maximizing the mutual information between classes, aligning with the objective of semi-supervised learning. Uncertainty-based selection provides a robust mechanism for acquiring pseudo-labels. The combination of the contrastive learning module and the uncertainty-based selection module forms a virtuous cycle to improve the performance of the proposed model. Extensive experiments demonstrate that ASCL outperforms state-of-the-art methods in terms of both convergence efficiency and performance. In the experimental scenario where only one label is assigned per class in the CIFAR-10 dataset, the application of ASCL to Pseudo-label, UDA (Unsupervised Data Augmentation for Consistency Training), and Fixmatch benefits substantial improvements in classification accuracy. Specifically, the results demonstrate notable improvements in respect of 16.32%, 6.9%, and 24.43% when compared to the original outcomes. Moreover, the required training time is reduced by almost 50%. © 2024, The Authors. All rights reserved.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

Skeleton-prompt: A cross-dataset transfer learning approach for skeleton action recognition

引用

Pattern Recognition 2026年 169卷

作者： Mingqi Lu Xiaobo Lu Jun Liu School of Automation Southeast University Nanjing 210096 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing 210096 China School of Computing and Communications Lancaster University UK

This paper presents Skeleton-Prompt, a novel tuning method designed to tackle cross-dataset transfer issues in skeleton action recognition models. Given the scarcity of large-scale 3D skeleton datasets and the variability in keypoint structures across datasets, existing methods often rely on training models from scratch, necessitating extensive labeled data and exhibiting high sensitivity to occlusion. Our approach aims to fine-tune pre-trained models to adapt to limited real-world skeleton data. We use 2D skeletons as inputs and leverage a large human motion dataset for 2D to 3D pose estimation to learn generalizable motion features. A lightweight prompt generator produces instance-level prompts, and we employ dynamic queries with cross-attention to refine the semantic information of the input data. Additionally, we introduce a joint-enhanced multi-stream fusion mechanism based on self-attention to improve robustness against incomplete skeletons. Skeleton-Prompt represents a significant advancement in efficient fine-tuning for skeleton action recognition, effectively addressing cross-dataset generalization challenges in a data-efficient and parameter-efficient manner.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Network compression via central filter

arXiv

引用

arXiv 2021年

作者： Duan, Yuanzhi Hu, Xiaofang Zhou, Yue Liu, Qiang Duan, Shukai College of Artificial Intelligence Southwest University China College of Electronics and Information Engineering Southwest University China Brain-inspired Computing & Intelligent Control of Chongqing Key Lab China Harbin Institute of Technology China

Neural network pruning has remarkable performance for reducing the complexity of deep network models. Recent network pruning methods usually focused on removing unimportant or redundant filters in the network. In this paper, by exploring the similarities between feature maps, we propose a novel filter pruning method-Central Filter (CF), which suggests that a filter is approximately equal to a set of other filters after appropriate adjustments. Our method is based on the discovery that the average similarity between feature maps changes very little, regardless of the number of input images. Based on this finding, we establish similarity graphs on feature maps and calculate the closeness centrality of each node to select the Central Filter. Moreover, we design a method to directly adjust weights in the next layer corresponding to the Central Filter, effectively minimizing the error caused by pruning. Through experiments on various benchmark networks and datasets, CF yields state-of-the-art performance. For example, with ResNet-56, CF reduces approximately 39.7% of FLOPs by removing 47.1% of the parameters, with even 0.33% accuracy improvement on CIFAR-10. With GoogLeNet, CF reduces approximately 63.2% of FLOPs by removing 55.6% of the parameters, with only a small loss of 0.35% in top-1 accuracy on CIFAR-10. With ResNet-50, CF reduces approximately 47.9% of FLOPs by removing 36.9% of the parameters, with only a small loss of 1.07% in top-1 accuracy on ImageNet. The codes can be available at https://***/8ubpshLR23/Central-Filter. Copyright © 2021, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

A high-bandwidth compact micro/nano-positioning stage based on a double parallelogram compliant mechanism

引用

Smart Materials and Structures 2025年第6期34卷 065015-065015页

作者： Tao Zhang Wantao Zhu Donglong Bian Xingxiang Li Rui Ma Wanyun Ding Yuhang Wang Faculty of Mechanical and Electrical Engineering Northeast Forestry University Harbin Heilongjiang 150000 People’s Republic of China Faculty of Computing and Control Northeast Forestry University Harbin Heilongjiang 150000 People’s Republic of China

This paper introduces a novel piezoelectric actuator (PEA)-driven two-degree-of-freedom micro/nano-positioning stage for high-precision adjustments in small-space situations. To enhance stage compactness and dynamic performance while significantly reducing parasitic motions, the composite leaf-type double parallelogram mechanism and the leaf-type double parallelogram mechanism are presented for implementation in guided and decoupled mechanisms, respectively. The composite bridge arm structure and compact double parallelogram design enable the amplification mechanism to provide significant lateral rigidity and safeguard the *** orthogonal arrangement of the composite bridge type amplification mechanism and the dual parallelogram decoupling guiding mechanism are integrated to achieve full decoupling of the stage at both the input and output ends. The compliance matrix method and Lagrange's equation were employed to establish mathematical analytical models of the static and dynamic characteristics, which were subsequently validated through finite element analysis. The stage's static and dynamic performances were thoroughly assessed through experiments. The closed-loop control of the PEA utilizing integral strain gauges effectively eliminates the nonlinear hysteresis issue on the stage and exhibits proficient trajectory tracking capability. The experimental findings indicate that the motion range of the precision positioning stage is 51.71 × 49.56 µm², the output coupling displacement error is under 1%, the displacement resolution is 12 nm, and the natural frequencies of the X-axis and Y-axis are 429 Hz and 438 Hz, respectively.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FaSTrack:A Modular Framework for Real-Time Motion Planning and Guaranteed Safe Tracking

FaSTrack:A Modular Framework for Real-Time Motion Planning a...

引用

作者： Chen, Mo Herbert, Sylvia L. Hu, Haimin Pu, Ye Fisac, Jaime Fernandez Bansal, Somil Han, Soojean Tomlin, Claire J. School of Computing Science Simon Fraser University BurnabyBCV5A 1S6 Canada Department of Mechanical and Aerospace Engineering University of California San Diego San DiegoCA92122 United States Department of Electrical Engineering Princeton University PrincetonNJ08536 United States Department of Electrical and Electronic Engineering University of Melbourne ParkvilleVIC3010 Australia Department of Electrical Engineering and Computer Sciences University of California Berkeley BerkeleyCA94704 United States Control and Dynamical Systems Program California Institute of Technology Los AngelesCA90019 United States

Real-time, guaranteed safe trajectory planning is vital for navigation in unknown environments. However, real-time navigation algorithms typically sacrifice robustness for computation speed. Alternatively, provably safe trajectory planning tends to be too computationally intensive for real-time replanning. We propose FaSTrack, Fast and Safe Tracking, a framework that achieves both real-time replanning and guaranteed safety. In this framework, real-time computation is achieved by allowing any trajectory planner to use a simplified planning model of the system. The plan is tracked by the system, represented by a more realistic, higher dimensional tracking model. We precompute the tracking error bound (TEB) due to mismatch between the two models and due to external disturbances. We also obtain the corresponding tracking controller used to stay within the TEB. The precomputation does not require prior knowledge of the environment. We demonstrate FaSTrack using Hamilton-Jacobi reachability for precomputation and three different real-time trajectory planners with three different tracking-planning model pairs. © 1963-2012 IEEE.

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

Numerical Simulation of Lowering a Payload into Waves with a Heave Compensation System Using Openfoam

SSRN

引用

SSRN 2022年

作者： Wang, Kai Ma, Xin Bai, Wei Qian, Ling Lin, Zaibin School of Control Science and Engineering Shandong University Shandong China Centre for Mathematical Modelling and Flow Analysis Department of Computing and Mathematics Manchester Metropolitan University ManchesterM1 5GD United Kingdom School of Engineering University of Aberdeen AberdeenAB24 3UE United Kingdom 17923 Jingshi Road. Shandong Jinan250061 China

In heavy lifting and lowering marine operations, heave compensation is wildly used to decouple the vessel heave motion from the payload and reduce the dynamic cable tension. This paper proposes a novel OpenFOAM-based numerical model to study the process of lowering payloads into waves by an offshore crane with a heave compensation system. First, an elastic cable module and electric winch system are established in a three-dimensional Numerical Wave Tank (NWT) model to achieve the crane vessel lifting and lowering operation. A correction method for the cable spring force is proposed to reduce the calculation error of the cable system. Second, a heave compensation is designed to compensate the vessel motion while taking hydrodynamic forces on payloads during the lowering operation into account. A physical experiment of lowering an aluminum cuboid into a water tank by the crane platform is conducted to verify the developed model. Finally, the process of lowering payloads into waves is simulated to analyze the dynamic character of the cable force and the payload motion using the developed numerical model. A PID controller is applied to stabilize the payload motion by controlling the winch velocity, reducing the dynamic cable tension and the risk of cable slack. © 2022, The Authors. All rights reserved.

关键词： Numerical models

来源：评论

学校读者我要写书评

暂无评论

Federated Coordinate Descent for Privacy-Preserving Multiparty Linear Regression

arXiv

引用

arXiv 2022年

作者： Leng, Xinlin Li, Chenxu Xu, Weifeng Sun, Yuyan Wang, Hongtao School of Control and Computer Engineering North China Electric Power University Baoding071051 China School of Computer Science and Technology Shandong University of Finance and Economics Jinan250014 China Beijing Key Laboratory of IoT Information Security Institute of Information Engineering Chinese Academy of Sciences Beijing100093 China Hebei Key Laboratory of Knowledge Computing for Energy & Power Baoding071051 China

Federated linear regressions have been developed and applied in various domains, where multiparties collaboratively and securely perform optimization algorithms, e.g., Gradient Descent, to learn a set of optimal model weights. However, traditional Gradient-Descent based methods fail to solve problems which contain L1regularization in objective functions, such as lasso regression. In this paper, we present Federated Coordinate Descent, a new distributed scheme called FCD, to address this issue securely under multiparty scenarios. Specifically, through secure aggregation and noise perturbations, our scheme guarantees that: (1) no local information is leaked to other parties, and (2) global model weights are not exposed to cloud servers. The proposed FCD scheme fills the gap between multiparty secure learning and coordinate descent methods, and is applicable for general linear regressions, including linear, ridge, and lasso regressions. Theoretical analyses prove the security guarantees of our scheme against semi-honest attackers. Experimental results on three UCI datasets demonstrate that regressions under our FCD scheme are effective as centralized models. The performance evaluation on synthetic dataset also reveals that our scheme achieves efficient properties on computation cost and communication *** Codes 68Q32 Copyright © 2022, The Authors. All rights reserved.

关键词： Gradient methods

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Generation of Tradable Topic Indices Through Textual Analysis

SSRN

引用

SSRN 2024年

作者： Lee, Marcel Spark, Alan Shanghai China Department of Economics Mathematics and Statistics Birkbeck College University of London Malet Street Bloomsbury LondonWC1E 7HX United Kingdom Department of Computer Science and Technology Tsinghua University Haidian District Beijing100084 China Department of Computing and Control Engineering University of Chemistry and Technology Technicka 5 Prague 6166 28 Czech Republic McKinsey & Company. Digital and Analytics Budapester Str. 46 Berlin10787 Germany

Stocks returns are moved by many risk factors. Thematic stock indices try to represent these factors, but are limited by the fact that risk factors are not directly observable. This paper shows a method to uncover hidden risk factors through text analysis. It applies the dynamic variant of the Latent Dirichlet Allocation (LDA) model to annual and quarterly reports to find a topic distribution for each stock. This is then interpreted as the risk factor partition and transformed into a standard normal basis which corresponds to pure risk factors. The weights indicate the proportions necessary to combine the equities into tradable topic indices. The need for human intervention is minimized by determining the optimal parameters automatically. © 2024, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Depth-Aware Multi-Modal Fusion for Generalized Zero-Shot Learning

Depth-Aware Multi-Modal Fusion for Generalized Zero-Shot Lea...

引用

IEEE International Conference on Industrial Informatics (INDIN)

作者： Weipeng Cao Xuyang Yao Zhiwu Xu Yinghui Pan Yixuan Sun Dachuan Li Bohua Qiu Muheng Wei Guangdong Laboratory of Artificial Intelligence and Digital Economy (Shenzhen) Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Stony Brook University New York United States Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Shenzhen China Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China ZhenDui Industry Artificial Intelligence Co. Ltd Shenzhen China Department of Automation Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China

ISBN: (数字)9798331527471

ISBN: (纸本)9798331527488

Realizing Generalized Zero-Shot Learning (GZSL) based on large models is emerging as a prevailing trend. However, most existing methods merely regard large models as black boxes, solely leveraging the features output by the final layer while disregarding potential performance enhancements from other layers. Indeed, numerous researchers have visually depicted variations in the features learned across different layers of neural networks. Motivated by this observation, we propose a Vision Transformer (ViT)-based GZSL method named Depth-Aware Multi-Modal ViT (DAM2ViT), which exploits multi-level features of ViT. DAM2ViT incorporates a multi-modal interaction block to align semantic information of categories across multiple layers, thereby augmenting the model's capacity to learn associations between visual and semantic spaces. Extensive experiments conducted on three benchmark datasets (i.e., CUB, SUN, AWA2) have showcased that DAM2ViT achieves competitive results compared to state-of-the-art methods.

关键词： Visualization Adaptation models Semantics Zero shot learning Neural networks Termination of employment Transformers Market research Sun Optimization

来源：评论

学校读者我要写书评

暂无评论

Car Plate Detection Based on Yolov3 5

Car Plate Detection Based on Yolov3

引用

2020 5th International Conference on Intelligent computing and Signal Processing, ICSP 2020

作者： Shen, Siyuan Wang, Lidan Duan, Shukai He, Xin College of Electronic and Information Engineering Southwest University Chongqing400715 China College of Artificial Intelligence Southwest University Chongqing400715 China Brain-inspired Computing and Intelligent Control of Chongqing Key Lab Chongqing400715 China National and Local Joint Engineering Laboratory Intelligent Transmission and Control Technology Chongqing400715 China Chongqing Brain Science Collaborative Innovation Center Chongqing400715 China

An intelligent car plate detection method can make the travel more convenient and efficient. However, traditional methods are reasonably effective under the specific circumstances or strong assumptions only, and there are few databases for car plate detection. Therefore, a novel real-time car plate detection method based on improved Yolov3 has been proposed. In order to select the more precise number of candidate anchor boxed and aspect ratio dimensions, the K-Means algorithm is utilized. To solve the short of the available car plate database, a car plate database which has 6668 pictures has been established. As shown in the experimental results, the method which is proposed by this paper is better than original Yolov3. Thanks to the car plate database, the proposed method obtained better results even in the situation of inclination, too bright or too dark, different weather and so on. © 2019 Published under licence by IOP Publishing Ltd.

关键词： Database systems

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：