检索结果-内蒙古大学图书馆

Parameter training methods for convolutional neural networks with adaptive adjustment method based on Caputo fractional-order differences

引用

Chaos, Solitons and Fractals 2025年 199卷

作者： Zhao, Haiming Yang, Honggang Chen, Jiejie Jiang, Ping Zeng, Zhigang School of Computer and Information Engineering Hubei Normal University Huangshi435000 China School of Automation Wuhan University of Technology Wuhan430000 China School of Computer Hubei Polytechnic University Huangshi435000 China School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430000 China Key Laboratory of Image Information Processing and Intelligent Control Ministry of Education of China Wuhan430000 China

As deep learning technologies continue to permeate various sectors, optimization algorithms have become increasingly crucial in neural network training. This paper introduces two adaptive momentum algorithms based on Grünwald–Letnikov and Caputo fractional-order differences—Fractional Order Adagrad (FAdagrad) and Fractional Order Adam (FAdam)—to update parameters more flexibly by adjusting momentum information. Commencing from the definitions of fractional derivatives, we propose integrating fractional-order differences with gradient algorithms in convolutional neural networks (CNNs). These adaptive momentum algorithms, leveraging Grünwald–Letnikov and Caputo fractional-order differences, offer enhanced flexibility, thereby accelerating convergence. Our nonlinear parameter tuning method for CNNs demonstrates superior performance compared to traditional integer-order momentum algorithms and the standard Adam algorithm. Experimental results on the BraTS2021 dataset and CIFAR-100 dataset reveal that the proposed fractional-order optimization algorithms significantly outperform their integer-order counterparts in model optimization. They not only expedite convergence but also improve the accuracy of image recognition and segmentation. © 2025 Elsevier Ltd

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Scaling properties of scale-free networks in degree-thresholding renormalization flows

arXiv

引用

arXiv 2021年

作者： Chen, Dan Cai, Defu Su, Housheng The School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China The Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Huazhong University of Science and Technology Wuhan430074 China The State Grid Hubei Electric Power Research Institute Wuhan China

We study the statistical properties of observables of scale-free networks in the degree-thresholding renormalization (DTR) flows. For BA scale-free networks with different sizes, we find that their structural and dynamical observables have similar scaling behavior in the DTR flow. The finite-size scaling analysis confirms this view and reveals a scaling function with a single scaling exponent that collectively captures the changes of these observables. Furthermore, for the scale-free network with a single initial size, we use its DTR snapshots as the original networks in the DTR flows, then perform a similar finite-size scaling analysis. Interestingly, the initial network and its snapshots share the same scaling exponent as the BA synthetic network. Our findings have important guiding significance for analyzing the structure and dynamic behavior of large-scale networks. Such as, in large-scale simulation scenarios with high time complexity, the DTR snapshot could serve as a substitute or guide for the initial network and then quickly explore the scaling behavior of initial networks. © 2021, CC BY.

关键词： Scaling laws

来源：评论

学校读者我要写书评

暂无评论

Dynamic Action Recognition Under Simulated Prosthetic Vision

Dynamic Action Recognition Under Simulated Prosthetic Vision

引用

International Conference on Networking and Network Applications (NaNA)

作者： Ying Zhao Dantong Xu Tie Wang Yanchun Ren Inner Mongolia Key Laboratory of Pattern Recognition and Intelligent Image Processing School of Information Engineering Inner Mongolia University of Science and Technology Baotou Inner Mongolia

ISBN: (数字)9781728189543

ISBN: (纸本)9781728189550

Purpose. To assess the recognition pattern of dynamic action by visual prosthesis' wearers. Methods. Twenty volunteers (classified by gender and experience) were recruited to carried out action recognition test in simulated prosthetic vision. Twenty-three common dynamic actions video clips usually see in daily life were selected and formed the experimental action set which was divided into three parts including combined actions, simple actions and difficult actions in turn. The action animations made by 3D Studio Max were recorded by Banicam to convert to MP4 video format and processed to phosphene video clips by MATLAB in different resolutions (16 × 16, 24 × 24, 32 × 32, 48 × 48, 64 × 64, 128 × 128). Result. Comparing the recognition results, male group and female group had no significant difference. 48 × 48 resolution was the considerable latent capacity which subjects had stable performance. After a period of study, the performance of prosthesis wearers could be improved significantly.

关键词： Three-dimensional displays Visual prosthesis Animation Pattern recognition Matlab

来源：评论

学校读者我要写书评

暂无评论

Coordinated operation of electricity and hydrogen blended gas networks via correlation-enhanced joint chance-constrained approach

引用

International Journal of Hydrogen Energy 2025年 149卷

作者： Sicheng Liu Bo Yang Lun Yang Xu Yang Xin Li Zhaojian Wang Kai Ma Xinping Guan State Key Laboratory of Submarine Geoscience School of Automation and Intelligent Sensing Shanghai Jiao Tong University Shanghai 200240 China Key Laboratory for System Control and Information Processing Ministry of Education of China Shanghai 200240 China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai 200240 China School of Automation Science and Engineering Xi’an Jiaotong University Xi’an 710049 China Key Lab of Industrial Computer Control Engineering of Hebei Province and the School of Electrical Engineering Yanshan University Qinhuangdao 066000 China

Blending green hydrogen into natural gas can effectively promote hydrogen consumption and operational flexibility of electricity-gas coupled networks. However, the volatility of renewable energy sources threatens the operation of green hydrogen production and blending, while existing system-level risk management methods either exhibit over-conservatism or are computationally inefficient when constraints are numerous. Therefore, this paper proposes a correlation-enhanced distributionally robust joint chance-constrained (DRJCC) method for the joint scheduling of electricity and hydrogen-enriched compressed natural gas (E-HCNG) networks, aiming to minimize operating cost while limiting joint violation risk under renewable energy uncertainty. Specifically, before DRJCC, an improved hydrogen fraction model is developed for hydrogen blending, refining the representation of fraction variations by incorporating the buffer effect of line packs. Then, for DRJCC solving, the commonly used Bonferroni approximations are overly conservative in allocating joint violation risks to individual chance constraints, especially when constraints are numerous. To mitigate this, by leveraging strongly violation correlations of operational constraints under affine policies, a grouping-based risk allocation strategy is proposed to reduce the over-conservatism. Theoretical derivations confirm the guarantee of joint violation risk via the proposed method. Additionally, customized transformations are developed to address the non-convexities introduced by variable hydrogen fractions. Case studies demonstrate the effectiveness and superiority of the proposed method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep dual support vector data description for anomaly detection on attributed networks

arXiv

引用

arXiv 2021年

作者： Zhang, Fengbin Fan, Haoyi Wang, Ruidong Li, Zuoyong Liang, Tiancai School of Computer Science and Technology Harbin University of Science and Technology Harbin China School of Information Engineering Zhengzhou University Zhengzhou China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Minjiang University Fuzhou China GRG Banking Technology Co. Ltd Guangzhou China

Networks are ubiquitous in the real world such as social networks and communication networks, and anomaly detection on networks aims at finding nodes whose structural or attributed patterns deviate significantly from the majority of reference nodes. However, most of the traditional anomaly detection methods neglect the relation structure information among data points and therefore cannot effectively generalize to the graph structure data. In this paper, we propose an end-to-end model of Deep Dual Support Vector Data description based Autoencoder (Dual-SVDAE) for anomaly detection on attributed networks, which considers both the structure and attribute for attributed networks. Specifically, Dual-SVDAE consists of a structure autoencoder and an attribute autoencoder to learn the latent representation of the node in the structure space and attribute space respectively. Then, a dual-hypersphere learning mechanism is imposed on them to learn two hyperspheres of normal nodes from the structure and attribute perspectives respectively. Moreover, to achieve joint learning between the structure and attribute of the network, we fuse the structure embedding and attribute embedding as the final input of the feature decoder to generate the node attribute. Finally, abnormal nodes can be detected by measuring the distance of nodes to the learned center of each hypersphere in the latent structure space and attribute space respectively. Extensive experiments on the real-world attributed networks show that Dual-SVDAE consistently outperforms the state-of-the-arts, which demonstrates the effectiveness of the proposed *** Codes 68T30 © 2021, CC BY.

关键词： Data description

来源：评论

学校读者我要写书评

暂无评论

One-shot distributed algorithm for PCA with RBF kernels

arXiv

引用

arXiv 2020年

作者： He, Fan Lv, Kexin Yang, Jie Huang, Xiaolin The Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China The MOE Key Laboratory of System Control and Information Processing 800 Dongchuan Road Shanghai200240 China

This letter proposes a one-shot algorithm for feature-distributed kernel PCA. Our algorithm is inspired by the dual relationship between sample-distributed and feature-distributed scenario. This interesting relationship makes it possible to establish distributed kernel PCA for feature-distributed cases from ideas in distributed PCA in sample-distributed scenario. In theoretical part, we analyze the approximation error for both linear and RBF kernels. The result suggests that when eigenvalues decay fast, the proposed algorithm gives high quality results with low communication cost. This result is also verified by numerical experiments, showing the effectiveness of our algorithm in practice. Copyright © 2020, The Authors. All rights reserved.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Customer churn warning with machine learning 5th

Customer churn warning with machine learning

引用

5th Euro-China Conference on intelligent Data Analysis and Applications, ECC 2018

作者： Wen, Zuotian Yan, Jiali Zhou, Liya Liu, Yanxun Zhu, Kebin Guo, Zhu Li, Yan Zhang, Fuquan Bank of China Beijing100091 China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Minjing University Fuzhou350121 China

ISBN: (纸本)9783030037659

Customer churn refers to the phenomenon of suspension of cooperation between customers and enterprises due to the implementation of various marketing methods. The customer churn warning refers to revealing the customer churn pattern hidden behind the data by analyzing the payment behavior, business behavior and basic attributes of the customer within a certain period of time, predicting the probability of the customer’s loss in the future and the possible reasons, and then guide the company to carry out customer retention work. After the forecast, the system can list the possible lost customers. And then the marketers can conduct precise marketing and improve marketing success rate. In this paper, we present a algorithm named Customer Churn Warning (CCW) to alert customers to churn. © Springer Nature Switzerland AG 2019.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

A UAV Path Planning Algorithm Based on an Improved D* Lite Algorithm for Forest Firefighting

A UAV Path Planning Algorithm Based on an Improved D* Lite A...

引用

Chinese Automation Congress (CAC)

作者： Zhaohui Luo Youmin Zhang Lingxia Mu Jing Huang Jing Xin Han Liu Shangbin Jiao Guo Xie Yingmin Yi Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing Xi’an University of Technology Xi’an China Concordia University Montreal Canada

ISBN: (数字)9781728176871

ISBN: (纸本)9781728176888

This paper proposes a new type of forest firefighting method. When one or more suspected fire spots are found in the forest, the firefighting Unmanned Aerial Vehicles (UAVs) can safely and quickly reach the desired location, conduct a thorough investigation of the suspected fire spots. If a fire occurs, it is necessary to drop firefighting materials to put out the fire. At the same time, considering the characteristics of limited endurance and low load capability have always limited the applications of UAVs, Unmanned Ground Vehicles (UGVs) are considered in this paper as an additional and comprehensive problem solution to provide energy and supplement materials for UAVs. In order to enable firefighting UAV to plan a safe and effective path in real time in the complex forest environment, a new improved D*Lite algorithm is proposed for UAVs in the mission team. Even when the 3D workspace is partially known or completely unknown, the algorithm can plan an effective path so that the firefighting UAV can move to the fire spots and UGV in real time.

关键词： Path planning Heuristic algorithms Unmanned aerial vehicles Forestry Smoothing methods Real-time systems Planning

来源：评论

学校读者我要写书评

暂无评论

MMD-RRT: a path planning strategy for robotic arm with improved RRT algorithm in unstructured environments

引用

Measurement Science and Technology 2025年第7期36卷 076203-076203页

作者： Peng Yang Hongbing Li Siqi Zhu Siyun Tan Yunpeng Lv Longzhou Cao Chongqing Key Laboratory of Geological Environment Monitoring and Disaster Warning in the Three Gorges Reservoir Area Chongqing Three Gorges University Chongqing 404120 People’s Republic of China Chongqing Key Laboratory of Intelligent Information Processing and Control Chongqing Three Gorges University Chongqing 404120 People’s Republic of China Internet of Things and Intelligent Control Technology Chongqing Engineering Research Center Chongqing Three Gorges University Chongqing 404120 People’s Republic of China

Robotic arm path planning in complex, unstructured environments often suffers from challenges such as excessive sampling randomness, low search efficiency, and redundant path nodes. To address these shortcomings, this paper introduces multi-mode dynamic sampling rapidly-exploring random tree (MMD-RRT), an enhanced path planning method based on the RRT algorithm. MMD-RRT dynamically adjusts its sampling strategy and step size in response to obstacles. key improvements include: (1) a novel multi-mode dynamic sampling strategy that effectively reduces the randomness of sample points; (2) a dynamic window with variable step sizes designed to adapt to diverse environments and improve search efficiency; and (3) path optimization using a greedy strategy to eliminate redundant nodes. Experimental results show that, compared to traditional algorithms in unstructured environments, MMD-RRT reduces path length by an average of 24.89% and search time by an average of 75.90%. Validation using an XArm6 robotic arm confirms the algorithm's ability to generate effective obstacle avoidance paths, ensuring the robot completes tasks safely and efficiently.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploiting learnable joint groups for hand pose estimation

arXiv

引用

arXiv 2020年

作者： Li, Moran Gao, Yuan Sang, Nong Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Tencent AI Lab United States

In this paper, we propose to estimate 3D hand pose by recovering the 3D coordinates of joints in a group-wise manner, where less-related joints are automatically categorized into different groups and exhibit different features. This is different from the previous methods where all the joints are considered holistically and share the same feature. The benefits of our method are illustrated by the principle of multi-task learning (MTL), i.e., by separating less-related joints into different groups (as different tasks), our method learns different features for each of them, therefore efficiently avoids the negative transfer (among less related tasks/groups of joints). The key of our method is a novel binary selector that automatically selects related joints into the same group. We implement such a selector with binary values stochastically sampled from a Concrete distribution, which is constructed using Gumbel softmax on trainable parameters. This enables us to preserve the differentiable property of the whole network. We further exploit features from those less-related groups by carrying out an additional feature fusing scheme among them, to learn more discriminative features. This is realized by implementing multiple 1x1 convolutions on the concatenated features, where each joint group contains a unique 1x1 convolution for feature fusion. The detailed ablation analysis and the extensive experiments on several benchmark datasets demonstrate the promising performance of the proposed method over the state-of-the-art (SOTA) methods. Besides, our method achieves top-1 among all the methods that do not exploit the dense 3D shape labels on the most recently released FreiHAND competition at the submission date. The source code and models are available at https://***/moranli-aca/LearnableGroups-Hand. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolution

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：