检索结果-内蒙古大学图书馆

Data Driven Control and Learning systems (DDCLS)

作者： Zhang Chen Wei Zhu Jingsui Li Hanlin Bian Chao Pei School of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing P. R. China Key Laboratory of Intelligent Analysis and Decision on Complex Systems Chongqing University of Posts and Telecommunications Chongqing P. R. China

ISBN: (数字)9798350361674

ISBN: (纸本)9798350361681

Neural networks are capable function approximators for data, but their lack of interpretability poses challenges. Physical-informed Polynomial Neural Ordinary Differential Equations (PIPNODEs) are proposed. They combine the constraints of prior knowledge of the data with interpretable Polynomial Neural Ordinary Differential Equations (PolyNODEs) to discover or approximate the governing equations. In this work, Neural Ordinary Differential Equations (NODEs) are categorized into two layers: a physical-informed term, which represents the priori knowledge of the data, and a discrepancy term, which is derived from the approximation of PolyNODEs. Pruning strategies are also used to remove redundant terms with coefficients lower than a certain threshold. We demonstrate the effectiveness and flexibility of this framework in a series of numerical algorithms with varying levels of prior knowledge. We found that having correct but incomplete physical information accelerates training and approaching thresholds compared to having incorrect physical information.

关键词： Training Accuracy Heuristic algorithms Ordinary differential equations Approximation algorithms Mathematical models Data models

来源：评论

学校读者我要写书评

暂无评论

Quantized Control for Highly Nonlinear Stochastic systems with Multiple Delays under DoS Attacks

Quantized Control for Highly Nonlinear Stochastic Systems wi...

引用

第42届中国控制会议

作者： Hairui Zhao Dongyan Chen Jun Hu School of Science Harbin University of Science and Technology Heilongjiang Provincial Key Laboratory of Optimization Control and Intelligent Analysis for Complex Systems Harbin University of Science and Technology School of Automation Harbin University of Science and Technology

In this paper,the quantized control problem is discussed for a class of highly nonlinear stochastic systems with multiple delays under the DoS *** coefficients are allowed to be highly nonlinear,and the control input is subject to the quantization effects and the DoS *** aim is to deal with the stabilization problem for unstable highly nonlinear stochastic systems with multiple *** p-th moment exponential stability and almost surely exponential stability are discussed in light of the Lyapunov ***,an illustrative example is given to verify the validity of the theoretical results.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DropPos: pre-training vision transformers by reconstructing dropped positions 23

DropPos: pre-training vision transformers by reconstructing ...

引用

Proceedings of the 37th International Conference on Neural Information Processing systems

作者： Haochen Wang Junsong Fan Yuxi Wang Kaiyou Song Tong Wang Zhaoxiang Zhang Center for Research on Intelligent Perception and Computing (CRIPAC) State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS) Institute of Automation Chinese Academy of Sciences (CASIA) and University of Chinese Academy of Sciences (UCAS) Center for Research on Intelligent Perception and Computing (CRIPAC) State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS) Institute of Automation Chinese Academy of Sciences (CASIA) and Centre for Artificial Intelligence and Robotics HKISI_CAS Megvii Technology Center for Research on Intelligent Perception and Computing (CRIPAC) State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS) Institute of Automation Chinese Academy of Sciences (CASIA) and University of Chinese Academy of Sciences (UCAS) and Centre for Artificial Intelligence and Robotics HKISI_CAS

As it is empirically observed that Vision Transformers (ViTs) are quite insensitive to the order of input tokens, the need for an appropriate self-supervised pretext task that enhances the location awareness of ViTs is becoming evident. To address this, we present DropPos, a novel pretext task designed to reconstruct Dropped Positions. The formulation of DropPos is simple: we first drop a large random subset of positional embeddings and then the model classifies the actual position for each non-overlapping patch among all possible positions solely based on their visual appearance. To avoid trivial solutions, we increase the difficulty of this task by keeping only a subset of patches visible. Additionally, considering there may be different patches with similar visual appearances, we propose position smoothing and attentive reconstruction strategies to relax this classification problem, since it is not necessary to reconstruct their exact positions in these cases. Empirical evaluations of DropPos show strong capabilities. DropPos outperforms supervised pre-training and achieves competitive results compared with state-of-the-art self-supervised alternatives on a wide range of downstream benchmarks. This suggests that explicitly encouraging spatial reasoning abilities, as DropPos does, indeed contributes to the improved location awareness of ViTs. The code is publicly available at https://***/Haochen-Wang409/DropPos.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Phenomenological Modeling and End-Point Trajectory Tracking Control of a Flexible Beam with Pneumatic Soft Actuators

Phenomenological Modeling and End-Point Trajectory Tracking ...

引用

IEEE International Conference of intelligent Robotic and Control Engineering (IRCE)

作者： Xuefeng Sun Qingxin Meng Jundong Wu Mingliang Zhu School of Automation China University of Geosciences Wuhan Hubei China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan Hubei China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan Hubei China

As the basic construction of flexible mechanical systems, the flexible beam and its control problem have attracted widespread attention in recent years. This paper takes a flexible beam with pneumatic soft actuators (FBPSA) as the research object and studies its phenomenological modeling and end-point trajectory tracking control strategy. Firstly, we develop an experimental platform for the FBPSA and perform tests on it, gathering its input and output data. Using these collected data, we analyzed the motion characteristics of the FBPSA. Subsequently, a phenomenological model is established to describe the motion characteristics of the system, and the parameters of this model are identified through a large amount of the collected experimental data. Then, a combined feedforward-feedback control strategy is proposed to achieve the end-point trajectory tracking control of the system. Finally, three sets of experiments are carried out to verify the accuracy of the established model and the effectiveness and superiority of the control strategy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Contribution-based Resource Allocation Scheme for Multi-population Methods in Dynamic Environments 2

A Contribution-based Resource Allocation Scheme for Multi-po...

引用

2nd International Conference on Industrial Artificial Intelligence, IAI 2020

作者： Peng, Mai Li, Changhe Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China

ISBN: (数字)9781728182162

ISBN: (纸本)9781728182162

The multi-population method is a common method for solving dynamic optimization problems. However, to design an efficient multi-population method, one of the challenging issues is how to allocate computational resources between populations given a limited computing buget in dynamic environments. This paper designs a contribution-based resource allocation mechanism. In this mechanism, a contribution degree of a population is defined according to the performance of the population, which determines the probability of the population to obtain the computing resource. This mechanism is implemented in an adaptive multi-population method. Experimental results on the moving peaks benchmark show that the algorithm equipped with the resource allocation mechanism outperforms the original algorithms. © 2020 IEEE.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Speech Emotion Recognition Based on Syllable-Level Feature Extraction

SSRN

引用

SSRN 2023年

作者： Rehman, Abdul Liu, Zhen-Tao Wu, Min Cao, Wei-Hua Jiang, Cheng-Shan School of Automation China University of Geosciences Wuhan430074 China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan430074 China

Speech emotion recognition systems have high computational requirements for deep learning models and low generalizability mainly because of the poor reliability of emotional measurements across multiple corpora. To solve these problems, we present a speech emotion recognition system based on a reductionist approach of decomposing and analyzing syllable-level ***-spectrogram of an audio stream is decomposed into syllable-level components, which are then analyzed to extract statistical features. The proposed method uses formant attention, noise-gate filtering, and rolling normalization contexts to decrease contextual differences and increase focus the attention on the structure of a formant. A set of syllable-level formant features is extracted and fed into a single hidden layer neural network that makes predictions for each syllable as opposed to the conventional approach of using a sophisticated deep learner to make sentence-wide *** syllable level predictions help to lower the aggregated error in utterance level cross-corpus predictions. The experiments on IEMOCAP (IE), MSP-Improv (MI), and RAVDESS (RA) databases show that the method archives better than the state-of-the-art cross-corpus unweighted accuracy of 47.6% for IE to MI and 56.2% for MI to IE. © 2023, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

A Novel Ceramic Surface Based Constant Force Control Method for Flatness Inspection

SSRN

引用

SSRN 2023年

作者： Chen, Xin Gao, Xu Zhou, Yang Qin, Lang School of Automation China University of Geosciences Wuhan430074 China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan430074 China

A constant force control method of the robot based on a force sensor is proposed for the difficult point of high difficulty and low accuracy of ceramic surface flatness inspection. This paper proposed a method that combines a constant impedance controller for normal contact force with a posture compliance controller. The method decomposes force control and position control orthogonally to simplify the complex three-dimensional surface model of ceramics. The method solves the problem that the end-effector is difficult to be perpendicular to the contact surface due to the change of ceramic surface shape. Moreover, the proposed method can also detect whether the ceramic surface is flat or not by the change of the force sensor's indication and the change of the actual end position of the robot arm. Finally, the defect detection experiments are conducted on a plane and on a complex surface, respectively. The experiments demonstrate the adaptability and fast convergence of the method proposed in this paper, which meets the needs of high precision and fast response of industrial processes. © 2023, The Authors. All rights reserved.

关键词： Force control

来源：评论

学校读者我要写书评

暂无评论

Measures and Optimization for Robustness and Vulnerability in Disconnected Networks

arXiv

引用

arXiv 2023年

作者： Zhu, Liwang Bao, Qi Zhang, Zhongzhi The Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Fudan University Shanghai200433 China The Research Institute of Intelligent Complex Systems Fudan University Shanghai200433 China The Shanghai Engineering Research Institute of Blockchain Shanghai200433 China

The function or performance of a network is strongly dependent on its robustness, quantifying the ability of the network to continue functioning under perturbations. While a wide variety of robustness metrics have been proposed, they have their respective limitations. In this paper, we propose to use the forest index as a measure of network robustness, which overcomes the deficiencies of existing metrics. Using such a measure as an optimization criterion, we propose and study the problem of breaking down a network by attacking some key edges. We show that the objective function of the problem is monotonic but not submodular, which impose more challenging on the problem. We thus resort to greedy algorithms extended for non-submodular functions by iteratively deleting the most promising edges. We first propose a simple greedy algorithm with a proved bound for the approximation ratio and cubic-time complexity. To confront the computation challenge for large networks, we further propose an improved nearly-linear time greedy algorithm, which significantly speeds up the process for edge selection but sacrifices little accuracy. Extensive experimental results for a large set of real-world networks verify the effectiveness and efficiency of our algorithms, demonstrating that our algorithms outperform several baseline schemes. Copyright © 2023, The Authors. All rights reserved.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Dynamic Sensing Modeling of Dielectric Elastomer Sensors Based on GRU Neural Network

Dynamic Sensing Modeling of Dielectric Elastomer Sensors Bas...

引用

IEEE International Conference of intelligent Robotic and Control Engineering (IRCE)

作者： Mingyi Zhang Peng Huang Yawu Wang School of Automation China University of Geosciences Wuhan Hubei China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan Hubei China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan Hubei China

Dielectric elastomer sensor (DES) is a flexible sensor that can perform free bending deformation, thus it has broad application prospects in the fields of medical electronics, wearable devices, soft robots, etc. Previous studies have mostly focused on exploring the static sensing characteristics of the DES. Considering that the DES may be used in dynamic conditions, it is very meaningful to study its dynamic sensing characteristics to broaden its application ranges. In this paper, a dynamic sensing model of the DES is established based on the gate recurrent unit (GRU) neural network. Firstly, the structure of the DES and the construction of the experimental system are introduced. In addition, the dynamic sensing characteristics of the DES are analyzed by conducting several sets of experiments, which shows that the DES has significant rate-dependent hysteresis nonlinearities, multivalued mapping and memory characteristics. After that, the dynamic sensing model of the DES is built based on the GRU neural network to describe the above dynamic sensing characteristics. Next, the dynamic displacement and force sensing models of the DES are trained according to the dynamic displacement-capacitance and dynamic force-capacitance experimental data, respectively. Finally, several experiments are performed to verify the effectiveness and generalization ability of the established dynamic sensing model.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DMNER: Biomedical Named Entity Recognition by Detection and Matching

DMNER: Biomedical Named Entity Recognition by Detection and ...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Bian, Junyi Jiang, Rongze Zhai, Weiqi Huang, Tianyang Huang, Xiaodi Zhou, Hong Zhu, Shanfeng Fudan University School of Computer Science Shanghai China Fudan University Institute of Science and Technology for Brain-Inspired Intelligence Shanghai China Charles Sturt University School of Computing Mathematics and Engineering Nsw Australia Atypon Systems Llc United Kingdom Zhangjiang Fudan International Innovation Center Fudan University Institute of Science and Technology for Brain-Inspired Intelligence Moe Frontiers Center for Brain Science Shanghai Key Lab of Intelligent Information Processing Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence Shanghai China

ISBN: (纸本)9798350386226

Biomedical Named Entity Recognition (NER) is a crucial task in extracting information from biomedical texts. However, the diversity of professional terminology, semantic complexity, and the widespread presence of synonyms pose significant challenges. Traditional methods that rely on sequence labeling training datasets often struggle to handle these complexities. To address this, we introduce a novel framework for BioNER, termed DMNER, which leverages external knowledge and operates in two steps: entity boundary detection and entity category identification through matching. The core of DMNER is its second step, which determines the entity category by retrieving similar entities and their categories from a knowledge dictionary using semantic similarity matching. Our experiments on 10 biomedical datasets demonstrate that DMNER outperforms baselines across these tasks, proving its effectiveness and adaptability. DMNER is versatile and can be applied to various NER tasks, including supervised NER, distantly supervised NER, and NER on multiple datasets with disjoint label sets. The DMNER code is publicly available. © 2024 IEEE.

关键词： Semantics

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：