检索结果-内蒙古大学图书馆

25th IEEE China Conference on System Simulation Technology and its Application, CCSSTA 2024

作者： Tang, Lan Kong, Jie Xu, Liang Xi'an Shiyou University School of Computer Xi'an China Tianjin University of Technology Tianjin Key Laboratory for Control Theory & Applications in Complicated Systems School of Electrical Engineering and Automation Tianjin China

ISBN: (纸本)9798350366600

Electronic medical records and doctor-patient conversations contain a wealth of useful information, such as disease symptoms, drug names, and cure cycles. Traditional deep learning approaches utilize bidirectional recurrent neural networks to encode text, which cannot effectively capture long-distance dependencies in text. In order to identify medical entity types in various texts more accurately, a named entity recognition approach based on RoBERTa-WWM-IDCNN-CRF is proposed. This approach introduces the RoBERTa-WWM pre-trained model, in order to improve the overall understanding of long Chinese texts for the first time. The model generates semantic representations containing prior knowledge. Then, the generated semantic representations are feed to the IDCNN to extract sentence level characteristics. Finally, the global optimal solution is obtained and the results of the named entity are produced through the CRF. The experiment indicates that the improvement of F1 score, accuracy, and recall of RoBERTa-WWM-IDCNN-CRF model are 7.94%, 4.52% and 10.02% comparing with the classical BiLSTM-CRF model. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Study of the Relationship Between Temperature and Efficiency in Photovoltaic Systems

Renewable Energy and Power Quality Journal

引用

Renewable Energy and Power Quality Journal 2023年 21卷 560-564页

作者： Salazar, Andres O. Fonseca, Diego A. de M. Ribeiro, Caio M. L. Xavier, Vinicius M. Lopes, Gabriela M. S. Maciel, Kalynne G. da S. Department Computer Engineering and Automation Federal University of Rio Grande do Norte Natal Brazil School of Science and Technology Federal University of Rio Grande do Norte Natal Brazil Department of Electrical Engineering Federal University of Rio Grande do Norte Natal Brazil

Photovoltaic solar energy has been growing due to the demand for diversification of the global energy grid and the concern about replacing fossil fuels with renewable energy. There is a great potential for this form of energy generation due to high levels of solar irradiation. However, the efficiency of commercial panels is still low and is negatively affected by temperature rises on the panel. To mitigate this effect, several cooling techniques have been previously proposed. This work focuses on studying the relation between the temperature of solar panels and the energy produced by them. It also uses thermoelectric cells acting as coolers for the system bringing down the panel temperature to increase its efficiency. The main goal is to build a plant consisting of a solar panel attached to one or more thermoelectric cells and connected to a data acquisition system. This setup will be used to evaluate the improvement in performance obtained through this strategy. The experiments will be carried out at the Petroleum Measurement Evaluation Laboratory-LAMP, and the results will be compared with mathematical and simulation models present in the literature. © 2023, European Association for the Development of Renewable Energy, Environment and Power Quality (EA4EPQ). All rights reserved.

关键词： Fossil fuels

来源：评论

学校读者我要写书评

暂无评论

Deep Security Detection Framework Based on ATT&CK

引用

International Journal of Network Security 2023年第1期25卷 161-170页

作者： Qiu, Rixuan Fu, Yu Le, Jian Zheng, Fuyong Qi, Gan Peng, Chao Zhang, Yi Li, Yuancheng Liu, Yan Dispatch Monitoring Center State Grid Jiangxi Information & Telecommunication Company Nanchang China School of Electrical Engineering and Automation Wuhan University Wuhan430072 China School of Control and Computer Engineering & North China Electric Power University Beijing102206 China

It is urgent to establish an efficient and comprehensive security system architecture facing the increasingly severe network threat. Aiming at the security problems in information systems, a deep security detection framework based on ATT&CK is proposed. There are three modules: information collection, analysis engine, and security response. The anomaly detection results of the technical layer based on ATT&CK knowledge base association analysis are mapped to the attack links of the tactical layer, thus realizing the bottom-up system security detection. An improved transformer anomaly detection method based on FLOATER positional coding is applied in the security detection framework to realize the early detection of network threats. Based on the continuous dynamic system, the positional coding function is learned to capture the temporal features of the sequence. Then multi-head attention is used to encode the input sequence. Moreover, the model training is optimized by the adversarial method. The experimental results demonstrated that the method applied in this work outperforms traditional deep learning models in terms of detection accuracy and F1 score. © 2022,International Journal of Network Security. All Rights Reserved.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Revisiting the Power of Prompt for Visual Tuning 41

Revisiting the Power of Prompt for Visual Tuning

引用

41st International Conference on Machine Learning, ICML 2024

作者： Wang, Yuzhu Cheng, Lechao Fang, Chaowei Zhang, Dingwen Duan, Manni Wang, Meng Zhejiang Lab China School of Computer Science and Information Engineering Hefei University of Technology China School of Artificial Intelligence Xidian University China School of Automation Northwestern Polytechnical University China

Visual prompt tuning (VPT) is a promising solution incorporating learnable prompt tokens to customize pre-trained models for downstream tasks. However, VPT and its variants often encounter challenges like prompt initialization, prompt length, and subpar performance in self-supervised pretraining, hindering successful contextual adaptation. This study commences by exploring the correlation evolvement between prompts and patch tokens during proficient training. Inspired by the observation that the prompt tokens tend to share high mutual information with patch tokens, we propose initializing prompts with downstream token prototypes. The strategic initialization, a stand-in for the previous initialization, substantially improves performance. To refine further, we optimize token construction with a streamlined pipeline that maintains excellent performance with almost no increase in computational expenses compared to VPT. Exhaustive experiments show our proposed approach outperforms existing methods by a remarkable margin. For instance, after MAE pre-training, our method improves accuracy by up to 10%∼30% compared to VPT, and outperforms Full fine-tuning 19 out of 24 cases while using less than 0.4% of learnable parameters. Besides, the experimental results demonstrate the proposed SPT is robust to prompt lengths and scales well with model capacity and training data size. We finally provide an insightful exploration into the amount of target data facilitating the adaptation of pre-trained models to downstream tasks. The code is available at https://***/WangYZ1608/Self-PromptTuning. Copyright 2024 by the author(s)

关键词： Pipeline codes

来源：评论

学校读者我要写书评

暂无评论

Quasi-Z-Source Cascaded H-Bridge Multilevel Inverters for PV Applications with Fault

Quasi-Z-Source Cascaded H-Bridge Multilevel Inverters for PV...

引用

2023 IEEE International Conference on Environment and Electrical engineering and 2023 IEEE Industrial and Commercial Power Systems Europe, EEEIC / I and CPS Europe 2023

作者： Carrasco-González, David Horrillo-Quintero, Pablo García-Triviño, Pablo Sarrias-Mena, Raúl García-Vázquez, Carlos Andrés Fernández-Ramirez, Luis M. University of Cadiz SURET Research Group Department of Electrical Engineering Algeciras Spain University of Cadiz SURET Research Group Department of Engineering in Automation Electronics and Computer Architecture and Networks Algeciras Spain

ISBN: (纸本)9798350347432

Nowadays, the use of quasi-Z-source cascaded H-bridge multilevel inverters (qZS-CHBMLI) in photovoltaic (PV) has gained their attention. They permit the single-stage power transformation of a PV generator adjusting the DC voltage independently in each module to operate each PV generator at maximum power. This paper develops a new control strategy that allows the system to continue operating after a fault in a module, by increasing the DC voltage in the operating modules. The control strategy is employed to a grid-connected qZS-CHBMLI, which consists of three cascaded qZSI powered by independent 4.8 kW PV cells. Under several operating circumstances, the proposed control system is evaluated. The results demonstrate how the control strategy applied to the system successfully reacts to the failure of a module. © 2023 IEEE.

关键词： Control systems

来源：评论

学校读者我要写书评

暂无评论

BOOSTED CURRICULUM REINFORCEMENT LEARNING 10

BOOSTED CURRICULUM REINFORCEMENT LEARNING

引用

10th International Conference on Learning Representations, ICLR 2022

作者： Klink, Pascal D'Eramo, Carlo Peters, Jan Pajarinen, Joni Department of Computer Science TU Darmstadt Germany Department of Electrical Engineering and Automation Aalto University Finland

Curriculum value-based reinforcement learning (RL) solves a complex target task by reusing action-values across a tailored sequence of related tasks of increasing difficulty. However, finding an exact way of reusing action-values in this setting is still a poorly understood problem. In this paper, we introduce the concept of boosting to curriculum value-based RL, by approximating the action-value function as a sum of residuals trained on each task. This approach, which we refer to as boosted curriculum reinforcement learning (BCRL), has the benefit of naturally increasing the representativeness of the functional space by adding a new residual each time a new task is presented. This procedure allows reusing previous action-values while promoting expressiveness of the action-value function. We theoretically study BCRL as an approximate value iteration algorithm, discussing advantages over regular curriculum RL in terms of approximation accuracy and convergence to the optimal action-value function. Finally, we provide detailed empirical evidence of the benefits of BCRL in problems requiring curricula for accurate action-value estimation and targeted exploration. © 2022 ICLR 2022 - 10th International Conference on Learning Representationss. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Customer Churn Prediction based on Stacking Model 4

Customer Churn Prediction based on Stacking Model

引用

4th International Conference on computer Vision, Image and Deep Learning, CVIDL 2023

作者： Chen, Fulu Wei, Xiaowei Yu, Saisai Ma, Pengfei He, Shuqing School of Information Science and Engineering Linyi University Linyi China School of Automation and Electrical Engineering Linyi University Linyi China School of Computer Science Qufu Normal University Rizhao China

ISBN: (纸本)9798350326444

In the process of product marketing, companies are likely to face the serious problem of losing their stock of customers. However, analyzing churned customers through data mining techniques can not only retain those customers who may be lost but also help companies target product optimization and management improvement. In this paper, a bank credit card customer churn prediction model is built using the Stacking model to predict the churn of credit card customers. The model has higher accuracy than XGBoost, random forest, and logistic regression models, runs well, and has better customer management implications. © 2023 IEEE.

关键词： Sales

来源：评论

学校读者我要写书评

暂无评论

New Skills and Knowledge for Digital Entrepreneurs in the age of Artificial Intelligence 19

New Skills and Knowledge for Digital Entrepreneurs in the ag...

引用

19th European Conference on Innovation and Entrepreneurship, ECIE 2024

作者： Orozova, Daniela Angelova, Nadezhda Zlatev, Zlatin Department of Informatics and Mathematics Faculty of Economics Trakia University Stara Zagora Bulgaria Department of Electrical Engineering Automation Computer Systems and Communications Faculty of Technics and Technologies Trakia University Yambol Bulgaria

ISBN: (纸本)9781917204149

Digital technologies are becoming increasingly complex and integrated, leading to significant transformations in society and the economy. The article aims to explore and summarize the new opportunities and potential risks of the widespread use of artificial intelligence (AI) in all aspects of life, to define the new skills and necessary knowledge of digital entrepreneurs and to highlight the need for transformation in modern education. Recognizing that the relationship between technology and business is two-way and becoming stronger, revealing that well-prepared employees are a guarantee of success and prosperity of companies in various fields, we try to focus on the main groups of qualities, skills and basic knowledge of students in the age of artificial intelligence. The development of the Internet, expansion of connectivity through social networks, the advent of AI, 3D printing, and immersive technologies like Augmented Reality, and Virtual Reality, require new knowledge and skills, leading to new challenges in education. Qualified personnel in this modern world must have solid professional training and systemic thinking (knowledge, skills, accumulated information), developed cognitive abilities, and personal skills based on collecting and analyzing large amounts of diverse information from heterogeneous sources. Questions arise: how can multiple information sources be combined effectively, and how can the fusion of multiple sources provide additional information to support decision-making processes? Combining information obtained from the real world makes the results heterogeneous and more informative. It follows the need to develop machine learning methods to extract relevant information from increasingly complex data sets. The goal is to improve the accuracy of the applied classification algorithms by combining predictions from multiple models, as well as obtaining a more stable final classification evaluation, effective handling of noisy data, adaptation to c

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Computing Performance Analysis of IEC 61499 Distributed automation Systems 26

Computing Performance Analysis of IEC 61499 Distributed Auto...

引用

26th International Conference on Industrial Technology, ICIT 2025

作者： Lyu, Tuojian Vyatkin, Valeriy Aalto University Department of Electrical Engineering and Automation Espoo Finland Luleå Tekniska Universitet Department of Computer Science Computer and Space Engineering Sweden

ISBN: (纸本)9798331521950

This work analyzes the computing performance of distributed control systems in IEC 61499 on two hardware devices, Dell XPS workstation and Raspi 4B. Based on the test IEC 61499 application, three different system configurations are tested to simulate the distribution options of singleprocess and single-thread, multi-threading, and multi-process on both hardware platforms. Experimental results show that multiprocess configurations yield more consistent execution time due to dedicated resource allocation, while multi-threading, though resource-efficient, can lead to latency variability. These results aim at help in choosing optimal configurations and selecting hardware for performance in distributed automation systems. © 2025 IEEE.

关键词： Parallel processing systems

来源：评论

学校读者我要写书评

暂无评论

Designed Steering Law for the Hybrid Actuators of Single Gimbal Control Moment Gyroscopes and Reaction Wheels Array 41

Designed Steering Law for the Hybrid Actuators of Single Gim...

引用

第41届中国控制会议

作者： Feng Liu Dongping Wang Kedong LYU School of electrical information engineering North Minzu University College of Computer and Information Engineering Tianjin Agricultural University Department of Automation Engineering Tianjin Electronic Information Technician College

ISBN: (数字)9789887581536

ISBN: (纸本)9781665482561

Single Gimbal Control Momentum Gyroscopes(SGCMG) are widely used with the ability of large torque output and high accuracy control of a *** singularity problem(especially the internal elliptical singularity) of SGCMG becoming a major obstacle in successful implementing the task of the attitude *** singularity can be effectively avoided by additional reaction wheel(RW) array in the field of the actual ***,the saturation problem is existed in the *** this paper,the hybrid actuators of the SGCMG with pyramid configuration and RW array with orthogonal configuration is *** new steering law is ***,the internal elliptical singularity and the saturation problem of the RW can be effectively avoided by the new steering law at the same time.

关键词： singularity avoidance new designed steering law hybrid actuators of SGCMG and RW array

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：