检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Zhu, Qunxi Guo, Yao Lin, Wei The Research Institute of Intelligent Complex Systems Fudan University Shanghai200433 China The Research Institute of Intelligent Complex Systems The Moe Frontiers Center for Brain Science State Key Laboratory of Medical Neurobiology The School of Mathematical Sciences Scms Scam Ccsb The Shanghai Artificial Intelligence Laboratory Fudan University Shanghai200433 China

Neural Ordinary Differential Equations (NODEs), a framework of continuous-depth neural networks, have been widely applied, showing exceptional efficacy in coping with representative datasets. Recently, an augmented framework has been developed to overcome some limitations that emerged in the application of the original framework. In this paper, we propose a new class of continuous-depth neural networks with delay, named Neural Delay Differential Equations (NDDEs). To compute the corresponding gradients, we use the adjoint sensitivity method to obtain the delayed dynamics of the adjoint. Differential equations with delays are typically seen as dynamical systems of infinite dimensions that possess more fruitful dynamics. Compared to NODEs, NDDEs have a stronger capacity for nonlinear representations. We use several illustrative examples to demonstrate this outstanding capacity. Firstly, we successfully model the delayed dynamics where the trajectories in the lower-dimensional phase space could be mutually intersected and even chaotic in a model-free or model-based manner. Traditional NODEs, without any argumentation, are not directly applicable to such modeling. Secondly, we achieve lower loss and higher accuracy not only for the data produced synthetically by complex models but also for the CIFAR10, a well-known image dataset. Our results on the NDDEs demonstrate that appropriately articulating the elements of dynamical systems into the network design is truly beneficial in promoting network *** Codes 34Kxx, 93C43, 82C32, 62M45, 68Txx Copyright © 2023, The Authors. All rights reserved.

关键词： Ordinary differential equations

来源：评论

学校读者我要写书评

暂无评论

Learning Transferable Negative Prompts for Out-of-Distribution Detection

arXiv

引用

arXiv 2024年

作者： Li, Tianqi Pang, Guansong Bai, Xiao Miao, Wenjun Zheng, Jin School of Computer Science and Engineering State Key Laboratory of Complex & Critical Software Environment Jiangxi Research Institute Beihang University China School of Computing and Information Systems Singapore Management University Singapore

Existing prompt learning methods have shown certain capabilities in Out-of-Distribution (OOD) detection, but the lack of OOD images in the target dataset in their training can lead to mismatches between OOD images and In-Distribution (ID) categories, resulting in a high false positive rate. To address this issue, we introduce a novel OOD detection method, named ‘NegPrompt’, to learn a set of negative prompts, each representing a negative connotation of a given class label, for delineating the boundaries between ID and OOD images. It learns such negative prompts with ID data only, without any reliance on external outlier data. Further, current methods assume the availability of samples of all ID classes, rendering them ineffective in open-vocabulary learning scenarios where the inference stage can contain novel ID classes not present during training. In contrast, our learned negative prompts are transferable to novel class labels. Experiments on various ImageNet benchmarks show that NegPrompt surpasses state-of-the-art prompt-learning-based OOD detection methods and maintains a consistent lead in hard OOD detection in closed- and open-vocabulary classification scenarios. Code is available at https://***/mala-lab/negprompt. © 2024, CC BY.

关键词： Chemical detection

来源：评论

学校读者我要写书评

暂无评论

Robust Delay-Dependent Stability for Uncertain Linear systems with Time-Varying Delay

Robust Delay-Dependent Stability for Uncertain Linear System...

引用

Annual Conference of Industrial Electronics Society

作者： Dong-Shuai Chen Yong He Xing-Chen Shangguan School of Automation China University of Geosciences Wuhan China Hubei key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education

This paper investigates the problem of stability analysis for the uncertain linear systems with time-varying delay. Firstly, an uncertain linear system model considering time-varying delay is established. Then based on the Lyapunov-Krasovskii functional (LKF) method, a novel robust delay-dependent stability criterion is proposed, which is benefited by a new augmented LKF with more effective time-delay information and the use of a tighter integral inequality to estimate functional derivative. The stability criterion obtained is less conservative. At last, a numerical example shows the superiority and effectiveness that the method used in this paper.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Stabilization of Stochastic Linear systems via Path-dependent Feedback Control

Stabilization of Stochastic Linear Systems via Path-dependen...

引用

Industrial Cyber-Physical systems (ICPS)

作者： Ruizhe Yu Xiaofeng Zong Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education School of Automation China University of Geosciences Wuhan China

This paper aims to investigate the stabilization problem of stochastic linear system via path-dependent state-feedback control. For the given stochastic linear system, a novel feedback control is designed with the path-dependent information of the system states, and the control gains are determined by the stochastic algebraic Riccati equation. To prove that path-dependent control can drive the stochastic linear system to be exponentially stable, a novel Lyapunov function is proposed. Combined with the general theory on stability of stochastic system, it is shown that stochastic system will be stabilized in mean-square via path-dependent control.

关键词： Linear systems Wireless sensor networks Stochastic systems Riccati equations Filling Feedback control Synchronization

来源：评论

学校读者我要写书评

暂无评论

Road Underground Target Detection in Ground Penetrating Radar Images Based on YOLO Models

Road Underground Target Detection in Ground Penetrating Rada...

引用

Chinese Automation Congress (CAC)

作者： Ting Li Wei Xue Xinya Yang Dongchang Wu Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education School of Automation China University of Geosciences Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan China

Ground penetrating radar (GPR) is extensively employed for subsurface road target detection, offering benefits such as convenience, nondestructive testing, rapid data acquisition, and superior resolution. Despite these advantages, interpreting GPR data often depends on the expertise of professionals, resulting in low detection efficiency and low accuracy. To address these challenges, this study introduces an intelligent detection technique for GPR images, utilizing an enhanced YOLOv5 framework. First, considering the problems of the small amount of GPR image datasets and the unclear characteristics caused by the complex underground media, a Dense-C3 module is built by utilizing the structure of DenseNet to enhance the network's capability for extracting features. Subsequently, a channel and spatial hybrid attention module is introduced into the backbone for feature refinement and improving the efficiency. Finally, the multi-class focal loss function is devised to enhance the precision in cases of imbalanced sample classes. Experimental results show that the proposed model surpasses the original YOLOv5 model and various contemporary advanced models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

USV Path Planning Under Marine Environment Simulation Using DWA and Safe Reinforcement Learning 19

USV Path Planning Under Marine Environment Simulation Using ...

引用

19th IEEE International Conference on Automation Science and Engineering, CASE 2023

作者： Qu, Tianci Xiong, Gang Ali, Hub Dong, Xisong Han, Yunjun Shen, Zhen Institute of Automation Chinese Academy of Sciences The State Key Laboratory of Multimodal Artificial Intelligence Systems Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China Institute of Automation Chinese Academy of Sciences Beijing Engineering Research Center of Intelligent Systems and Technology Beijing100190 China Chinese Academy of Sciences Guangdong Engineering Research Center of 3D Printing and Intelligent Manufacturing Cloud Computing Center Dongguan523808 China Qingdao Academy of Intelligent Industries Qingdao266109 China

ISBN: (纸本)9798350320695

It is a challenge for USV navigation due to uncertainties and disturbances in the complex marine environment, which may lead to tilting or collision. However, current path planning methods for USVs lack dynamic environment adaptivity and timeliness. We formulate this problem as a Markov Decision Process (MDP) and combine the advantages of Dynamic Window Approach (DWA) and Safe Reinforcement Learning (RL) to solve it. First, the state encoding output by DWA is employed as observation to reduce the state dimension. Second, a safety-ensured reward function is designed to avoid collision areas. Then, the actor-critic network based on RL algorithm is used. Along with the RL training, the parameters in DWA is adaptively adjusted. Simulation experiments with dynamic waves and ocean currents demonstrate the adaptivity of our method, with avg. trajectory length decreased by 10%, avg. time cost decreased by 17%, and avg. speed increased by 18%, compared to the case using only DWA. © 2023 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Hopf Bifurcation, Chaos and Bursting in a Generalized Logistic Model with Time Delays 2

Hopf Bifurcation, Chaos and Bursting in a Generalized Logist...

引用

2nd International Conference on Applied Mathematics, Modeling and Computer Simulation, AMMCS 2022

作者： Li, Shunyi Wan, Xiaoli Xiong, Mei School of Mathematics and Statistics Qiannan Normal University for Nationalities Duyun Guizhou558000 China Key Laboratory of Complex Systems and Intelligent Optimization of Guizhou Duyun Guizhou558000 China Qiannan Medical College for Nationalities Duyun Guizhou558000 China

ISBN: (纸本)9781643683522

A generalized Logistic system with two delays is investigated, and the conditions for the positive equilibrium occurring local Hopf bifurcation are obtained. Numerical results show that delayed system considered have complicated dynamic behaviours, such as period oscillation, doubling-bifurcations, chaos and Hopf bursting. © 2022 The authors and IOS Press.

关键词： Hopf bifurcation

来源：评论

学校读者我要写书评

暂无评论

Landslide Sensitivity Mapping Based on One-Dimensional Residual Convolutional Neural Networks

Landslide Sensitivity Mapping Based on One-Dimensional Resid...

引用

2022 Chinese Automation Congress, CAC 2022

作者： Ke, Yougang Lu, Chengda Wu, Min Yu, Wanke Niu, Ruiqing School of Automation China University of Geosciences Wuhan China Hubei Key Laboratory of Advanced Control Intelligent Automation for Complex Systems Wuhan China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan China Institute of Geophysics and Geomatics China University of Geosciences Wuhan China

ISBN: (纸本)9781665465335

Landslide is a common geological disaster. Landslide sensitivity mapping (LSM) is the key technology for landslide monitoring, early warning and risk assessment. Deep learning shows good performance in feature extraction. This paper proposes a one-dimensional residual convolution neural network (1DRCNN), which takes Yunyang County, Chongqing City, China, located in the Three Gorges Reservoir area as the research area. Extracting 12 evaluation factors from multi-source remote sensing data and building a training set. The spatial probability of landslide occurrence is quantitatively predicted by the proposed model, and finally, the landslide sensitivity mapping is generated. Compared with the common machine learning models SVM and logistic regression, the results show that the AUC (area under curve) and accuracy of 1D-RCNN are 0.9860 and 93.38%, respectively, which proving that this method is effective and can provide a reference for disaster prevention and reduction. © 2022 IEEE.

关键词： Landslides

来源：评论

学校读者我要写书评

暂无评论

A multiple-time-scale control strategy for burden distribution and blast supply of a BF

A multiple-time-scale control strategy for burden distributi...

引用

Chinese Control and Decision Conference, CCDC

作者： Yunpeng Guo Jianqi An Guoyu Zhao Min Wu Qingyi Wang School of Automation China University of Geosciences Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan China

The operations of blast furnaces (BFs) are very vital for the long-term stability of the iron making process. The burden distribution and blast supply are the two major operation systems of BFs. At present, the researches are lack effective adjustment methods for the BF operations combined with burden distribution and blast supply. The burden distribution affects the iron making process on a long time scale, while the blast supply affects the iron making process on a short time scale. This paper presents a multi-time sampling-data adjustment strategy for the BF operations aiming at optimizing GUR on multiple time scales. First, this paper analyzes the relationship between the gas utilization ratio (GUR) and the burden distribution, the blast supply on multiple time scales. Then, this paper establishes a prediction model of GUR on the long time scale and the short time scale based on autoregressive integrated moving average (ARIMA). Next, this paper provides a control strategy of burden distribution and a control strategy of blast supply by a prediction model based on support vector regression (SVR). Finally, this paper makes experiments and applies this method in a real-world BF. The analysis of the results shows the control strategy of the BF operations provides a good guide on making a suitable decision for burden distribution and blast supply.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Attention-Based Predictive Model in Multi-Agent Reinforcement Learning

Research on Attention-Based Predictive Model in Multi-Agent ...

引用

2022 Chinese Automation Congress, CAC 2022

作者： Hong, Weixin Wu, Huaiyu Shen, Zhen Xiong, Gang Hu, Bin Dong, Xisong Wang, Fei-Yue The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China The Beijing Engineering Research Center of Intelligent Systems and Technology Institute of Automation Chinese Academy of Sciences Beijing100190 China The Guangdong Engineering Research Center of 3D Printing and Intelligent Manufacturing Cloud Computing Center Chinese Academy of Sciences Dongguan523808 China

ISBN: (纸本)9781665465335

Agent actions planning is a challenging problem in multi-agent reinforcement learning. Recent methods typically build their predictive models by full connection layers, but the shortage of utilizing action and observation information in complex multi-agent environment limits the performance of those methods. To address this issue, we propose a novel predictive model based on self-attention mechanism which can effectively use information in the environment. Specifically, we use the MADDPG algorithm to make a dataset where the input is the observations and actions and the output is the predicted observations and rewards. We also introduce the multi-head attention mechanism into the predictive model so that the predictive model can better combine different information for prediction. We perform extensive experiments to assess the proposed method and compare it with the representative baseline approach. Results demonstrate that our method outperforms competitors with higher accuracy in different tasks. Our implementation will be publicly available. © 2022 IEEE.

关键词： Reinforcement learning

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：