检索结果-内蒙古大学图书馆

IEEE/RSJ International Conference on Intelligent Robots and systems (IROS)

作者： Zhi Li Jinghao Xin Ning Li Department of Automation Shanghai Jiao Tong University Shanghai P.R. China Ministry of Education of China Key Laboratory of System Control and Information Processing Shanghai China Shanghai Engi-neering Research Center of Intelligent Control and Management Shanghai China

Deep reinforcement learning (DRL) has been widely applied in autonomous exploration and mapping tasks, but often struggles with the challenges of sampling efficiency, poor adaptability to unknown map sizes, and slow simulation speed. To speed up convergence, we combine curriculum learning (CL) with DRL, and first propose a Cumulative Curriculum Reinforcement Learning (CCRL) training framework to alleviate the issue of catastrophic forgetting faced by general CL. Besides, we present a novel state representation, which considers a local egocentric map and a global exploration map resized to the fixed dimension, so as to flexibly adapt to environments with various sizes and shapes. Additionally, for facilitating the fast training of DRL models, we develop a lightweight grid-based simulator, which can substantially accelerate simulation compared to popular robot simulation platforms such as Gazebo. Based on the customized simulator, comprehensive experiments have been conducted, and the results show that the CCRL framework not only mitigates the catastrophic forgetting problem, but also improves the sample efficiency and generalization of DRL models, compared to general CL as well as without a curriculum. Our code is available at https://***/BeamanLi/CCRL_Exploration.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on NLOS environmental ranging method based on gradient boosting tree 8

Research on NLOS environmental ranging method based on gradi...

引用

8th International Conference on Electrical, Mechanical and Computer Engineering, ICEMCE 2024

作者： Li, Wenyu Cao, Shun Liu, Xianda Wang, Minghao Sheng, Shuan Shenyang University of Chemical Technology School of Information Engineering Liaoning Shenyang China Shenyang Institute of Automation Chinese Academy of Sciences Liaoning Shenyang China Chinese Academy of Sciences Key Laboratory of Networked Control System Liaoning Shenyang China Shenyang National Research Institute of Intelligent Robots Co. Ltd. Liaoning China

ISBN: (纸本)9798331506230

Indoor distance measurement plays a critical role in positioning services. The complexity of indoor environments, coupled with a limited and small sample size, results in suboptimal distance accuracy. This paper proposes a method for non-line-of-sight distance measurement based on a gradient boosting tree model, effectively addressing the issue of sample size dependency. The random hyperparameter settings of the gradient boosting tree model lead to prolonged fitting times. Additionally, the model utilizes a fixed learning rate, which limits the analysis of features in small samples. To tackle these concerns, we develop a grid optimization algorithm that identifies the best hyperparameter combinations through a hyperparameter grid search. We also implement a weighted value approach and a cosine annealing algorithm to precisely update the learning rate according to training conditions, thus maximizing the learning of small sample features. Our proposed model significantly enhances distance accuracy in non-line-of-sight environments under constraints of training sample size and variety. Comparative experiments demonstrate that our model improves distance accuracy by 63% relative to other machine learning models, with a mean square error of only 0.168m. © 2024 IEEE.

关键词： Distance measurement

来源：评论

学校读者我要写书评

暂无评论

Research on Named Entity Recognition for Oral History Text

Research on Named Entity Recognition for Oral History Text

引用

Culture-Oriented Science and Technology (CoST), International Conference on

作者： Chaoyu Hu Shilei Bai Mengfei Zhang School of Information and Communication Engineering Communication University of China Beijing China Beijing Key Laboratory of Modern Entertainment Technology Beijing China Key Laboratory of Acoustic Visual Technology and Intelligent Control System Ministry of Culture Beijing China

ISBN: (数字)9798350380347

ISBN: (纸本)9798350380354

Oral history involves conveying experienced events through spoken narratives, typically preserved and later studied by converting spoken language into text. Named entity recognition (NER) is the recognition of meaningful named entities in text, such as person names, locations, organizations, etc. Conducting NER research on oral history text is significant as it not only facilitates the verification of named entities by researchers but also allows for the extraction of key information from texts through named entities. By collecting oral history corpus, we establish an oral history text dataset, manually annotating all texts with six entity labels and repeatedly verifying them manually. Utilizing the concept of machine reading comprehension (MRC), a semantic fusion module was added, integrating label information into text information using a cross-attention mechanism, and decoding the corresponding entity labels through span decoding. The final model performs well on the oral history text dataset and also showed good performance on general Chinese datasets.

关键词： Costs Text recognition Semantics Named entity recognition Organizations Decoding History

来源：评论

学校读者我要写书评

暂无评论

Building Energy Optimization Based on Biased ReLU Neural Network

Building Energy Optimization Based on Biased ReLU Neural Net...

引用

第40届中国控制会议

作者： Hongyi Li Xinglong Liang Jun Xu Harbin Institute of Technology Key Laboratory of System Control and Information Processing Ministry of Education

This paper proposes a building energy optimization strategy based on artifical intelligence technology modeling ***,the data set generated by Energy Plus energy consumption simulation software is used as the training set and test set of the Biased ReLU neural network（BRNN）.Secondly,the building energy consumption prediction model and indoor temperature prediction model are built based on the Biased ReLU neural ***,model predictive control（MPC） is uesd to achieve energy saving by controlling the set temperature of the building’s Heating,Ventilation and Air Conditioning（HVAC） ***,the joint simulation of MATLAB and EnergyPlus is realized by introducing the building control virtual test bed（BCVTB）.The results show that our method can effectively reduce building energy consumption.

关键词： Biased ReLU neural network MPC building energy consumption HVAC system

来源：评论

学校读者我要写书评

暂无评论

Continual Semantic Segmentation via Mask-Based Class Rebalancing

Continual Semantic Segmentation via Mask-Based Class Rebalan...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Yongjie Guo Siya Chen Hongjian You Key Laboratory of Technology in Geo-Spatial Information Processing and Application System Aerospace Information Research Institute Chinese Academy of Sciences Beijing China School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Continual semantic segmentation (CSS) has risen as a popular field, which aims to acquire new skills constantly without forgetting past knowledge catastrophically. In CSS, we identify that there is a severe imbalance between new classes and old classes, leading to the classifier weight toward new classes. In this paper, we deal with the continual semantic segmentation problem from the class imbalance perspective via mask-based class rebalancing, avoiding the model suffering from catastrophic forgetting. More specifically, the mask-based class rebalancing depends on a mask to combine resampling with reweighting ingenuously, which mitigates the classifier bias toward new classes. Besides, we also propose a frequency knowledge distillation, leveraging multiple frequency components information to maintain the feature representation space for old classes. We demonstrate the effectiveness of our approach with an extensive evaluation of the Pascal-VOC 2012 and ADE20K datasets, significantly outperforming the state-of-the-art method.

关键词： Semantic segmentation Semantics

来源：评论

学校读者我要写书评

暂无评论

Remote Sensing Image Object Detection Based on Improved Sparse R-CNN

Journal of Network Intelligence

引用

Journal of Network Intelligence 2023年第4期8卷 1303-1320页

作者： Zhao, Li-Quan Chen, Chun-Lu Zhong, Tie Cui, Ying Jia, Yan-Fei Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology Ministry of Education Northeast Electric Power University Jilin 132012 China School of Electric Power Engineering Northeast Electric Power University Jilin 132012 China Zhuhai Power Supply Bureau Guangdong Electric Power Corporation Zhuhai 519000 China School of Electrical and Information Engineering Beihua University Jilin 132013 China

A more precise sparse region convolutional neural network algorithm is developed to increase the accuracy of remote sensing object recognition. Firstly, a new intersection over union method is proposed to solve the problem that the intersection over union used in sparse region convolutional neural network cannot well measure the offset between the ground-truth box and the proposal box. It introduces multiple geo-metric factors that are overlap area, center point distance, vertex distance between the ground-truth box and the proposal box, and diagonal distance of minimum enclosing of the ground-truth box and the proposal box and into the new intersection over union. The proposed intersection over union can measure the offset between the ground-truth box and the proposal box to optimal the proposal box. Secondly, to improve the detection precision of sparse region convolutional neural network, a dual branches dynamic instance interaction head is designed. The proposed dynamic interaction head consists of a fully connected branch and a convolution branch. The fully connected head is used for both bounding box regression and object classification. The convolution branch is only used for bounding box regression. Compared with original sparse region convolutional neural network, the average mAP and recall are 3.55% and 5.75% higher, and the average detection speed is only 0.85 FPS slower. The proposed method realizes balance between detection precision and detection speed. © 2023, Taiwan Ubiquitous information CO LTD. All rights reserved.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Robust sliding model predictive control with disturbance compensation in application to AUV pitch-axis motion

Robust sliding model predictive control with disturbance com...

引用

Chinese control and Decision Conference, CCDC

作者： Jiuwu Hui Jingqi Yuan State Key Laboratory of Nuclear Power Safety Monitoring Technology and Equipment Shenzhen Department of Automation and the Key Laboratory of System Control and Information Processing/Ministry of Education Shanghai Jiao Tong University Shanghai

ISBN: (纸本)9781665478977

Constraints on the control input and system states as well as uncertain disturbances are the practical challenges in many motion control systems. In this paper, a disturbance observer (DO) with two-step disturbance compensation algorithm is proposed to deal with disturbances and enhance the robustness of the controller, while a discrete sliding mode predictive control (DSMPC) scheme is designed to handle saturation constraints on the control input and system states with the application to an autonomous underwater vehicle (AUV). The DSMPC combines the model predictive control (MPC) with the discrete sliding mode control (DSMC) using linear matrix inequalities approach. The stability of the proposed DSMPC are proved via Lyapunov stability theorem. Finally, simulation results are presented to illustrate the effectiveness of the whole control scheme.

关键词： Simulation Prediction algorithms Stability analysis Robustness Linear matrix inequalities Motion measurement Motion control

来源：评论

学校读者我要写书评

暂无评论

Generating Molecules Conditional on 3D Protein Pockets with HGAF

Generating Molecules Conditional on 3D Protein Pockets with ...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Qinglian Zhu Xiaoli Lin Xiaolong Zhang Hubei Key Laboratory of Intelligent Information Processing and Real-Time Industrial System School of Computer Science and Technology Wuhan University of Science and Technology Wuhan China

Most current representations of protein pockets are the atom-pair graph, which ignore the global structural information of amino acids. Therefore, we propose a new molecular generation model, which uses the hypergraph to represent protein pocket structure, and combines the structural features obtained by atom-pair graph representation. These two levels of graphs are more capable of representing the complex structural information of protein pockets. Then, the graphs of the two levels of the protein pockets are input into the improved network model, which is named Hypergraph Graph Attention Fusion (HGAF), to obtain the embedding representation of the protein pockets, which is used as a condition to constrain the molecule generation. The molecules sampled by HGAF are subjected to quality assessment and docking targeting validation. Experimental results show that the molecules generated by proposed method can achieve better results in both of these assessment approaches.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Anti-3CLpro Molecular Design Based on the Model Constrained by Specific DTIs

Anti-3CLpro Molecular Design Based on the Model Constrained ...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Xiaoli Lin Xuan Liu Xiaolong Zhang Hubei Key Laboratory of Intelligent Information Processing and Real-Time Industrial System School of Computer Science and Technology Wuhan University of Science and Technology Wuhan China

Computer-aided drug design and artificial intelligence-driven drug design have accelerated drug discovery. However, how to design effective drugs that have strong interaction ability with target proteins to further improve the efficacy of drugs in treating diseases remains a key issue. This paper proposes a new target-specific drug generative model 3CLpro2mol to generate new drug molecules, which uses features of drug-target interactions (DTIs) to constrain the correlation between the drug and the target protein. To obtain as many drug-target interaction features as possible from a small amount of data, a small molecule extraction strategy is proposed to ensure the diversity of small molecules in the training samples. To improve the efficiency and accuracy of the generative model, a TOP K sampling strategy is used to generate tokens, which can improve the rationality and diversity of the generated molecules. The experimental results show that the proposed model has the potential to generate small molecules that interact better with the target protein.

关键词：

来源：评论

学校读者我要写书评

暂无评论

AoI-Reliability Analysis with Block Length over Erasure Channels Using Rateless Codes

AoI-Reliability Analysis with Block Length over Erasure Chan...

引用

International Conference on Communication Technology (ICCT)

作者： Haobo Huang Yusun Fu Yue Qiao Junpeng Yin Weiwu Yan Ningbo Artificial Intelligence Institute Shanghai Jiao Tong University Ningbo China School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

As a metric to measure the information freshness, Age of information (AoI) is widely used to evaluate low-latency high-reliability communication. Since the communication in Industrial Internet of Things (IIoT) often has the special requirements, the relationship among AoI, reliability and block length in time-critical scenarios with bounded latency and non-time-critical scenarios without bounded latency is explored respectively in this paper, based on erasure channels using rateless codes. We build a model using LT codes over erasure channels and consider two metrics: average AoI and bounded AoI. Theoretical analysis and simulation results show that AoI has a minimum value with different block lengths. Furthermore, we draw a conclusion that the minimum value of AoI under different reliability is monotonically decreasing functioned with reliability. Finally, the impact of overhead on reliability and AoI of LT codes in time-critical scenarios is studied.

关键词： Measurement Codes Simulation Reliability theory information age Communications technology Time factors Low latency communication Industrial Internet of Things

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：