检索结果-内蒙古大学图书馆

International Joint Conference on Neural Networks (IJCNN)

作者： Caidong Tan Xiao Sun AnHui Province Key Laboratory of Affective Computing and Advanced Intelligent Machine School of computer and information Hefei University of Technology Hefei China

Contrastive learning can reduce the impact of ex-posure bias associated with training using maximum likelihood estimation, which aims to pull together positive samples to increase the likelihood of high-quality summaries and push away irrelevant negative samples to reduce the likelihood of low-quality summaries. In contrastive learning-based text summarization methods, a standard method for selecting positive and negative samples is randomly selected within a batch. This method can lead to sampling bias to the extent that the consistency of the representation space is compromised. Therefore, we propose a new method to penalize false negatives based on ROUGE metric scores as weights to sample from the dynamic output of the model training process. The method calculates ROUGE metric scores for penalizing false negatives in real-time and can distinguish between positive and negative samples to ensure spatial consistency and alleviate exposure bias. Experimental results on XSum, CNN/DM, and Multi-News datasets show that our approach effectively improves the performance of the latest text summarization pre-training models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Structural Optimization of a Magnetically Operated Actuator Mechanism for 10kV Circuit Breaker 25

Structural Optimization of a Magnetically Operated Actuator ...

引用

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

作者： Liang, Chao Tian, He Zhao, Guanqiao Li, Jiangping Zhu, Ziliang Shen, Ruixuan Tianjin University of Technology Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control School of Mechanical Engineering Tianjin China Tianjin University of Technology National Demonstration Center for Experimental Mechanical and Electrical Engineering Education Tianjin China Tianjin Sanyuan Power Equipment Manufacturing Co. Ltd Tianjin China

ISBN: (纸本)9798350366600

As the demand for reliable power supply in power systems grows, ensuring the performance of circuit breaker operating mechanisms is crucial. This performance is pivotal to the stability of the power grid, the speed of fault recovery, and the equipment's lifespan. In this paper, we introduce a structural optimization plan for a new magnetic control operating mechanism designed for 10kV circuit breakers. Employing three-dimensional and mathematical modeling, complemented by finite element analysis with Maxwell software, we delved into an in-depth study of the mechanism's structure, performance, and materials. We further conducted a parametric design and comparative analysis of key components. The simulation results, which exclude the spring load, demonstrate that the optimized circuit breaker can achieve a closing velocity of 3.07 m/s, an electromagnetic force peaking at 2.7 kN, and a closure time minimized to 8.5 ms. These enhancements notably improve the mechanism's closing efficiency. Consequently, our research introduces a novel perspective to the field of reliability studies for magnetic control operating mechanisms. © 2024 IEEE.

关键词： Magnetic actuators

来源：评论

学校读者我要写书评

暂无评论

INNSE: Invertible neural network-based DNA image storage with self-correction encoding

引用

Computational and Structural Biotechnology Journal 2025年 27卷 2492-2502页

作者： Yanfen Zheng Xiaokang Zhang Lijun Sun Ben Cao Bin Wang Qiang Zhang Pan Zheng School of Computer Science and Technology Dalian University of Technology Dalian Liaoning 116024 China The Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education School of Software Engineering Dalian University Dalian Liaoning 116622 China Department of Accounting and Information Systems University of Canterbury Upper Riccarton Christchurch 8140 New Zealand

DNA storage has become an alternative to silicon-based storage of media. However, when used to store multimodal data such as images, it fails to take full advantage of the characteristics of high correlation and variability present in such data, and currently remains inferior in terms of encoding density and time complexity. It also has the shortcoming of weak performance in terms of error correction methods to address insertion and deletion errors caused by error crosstalk. Therefore, we proposed an invertible neural network-based DNA image storage self-correction encoding method (INNSE) that uses the invertibility of network to perform up-down sampling processing on image and video data, thereby greatly reducing the number of DNA sequences required for data storage. A codebook generator was then designed to encode low-dimensional vectors into DNA sequences that meet various constraints. Finally, a self-correction method was constructed to achieve DNA sequence error correction at the base level without adding any redundant information. Our experimental results showed that, compared to the Yin-Yang algorithm, the encoding density of INNSE was increased by 3×, encoding time complexity was reduced by 95.26%. Under a simulated error rate of 2%, PSNR and SSIM were raised by 80.88% and 63.82%, respectively, proving that INNSE has significant advantages in encoding performance and error correction ability.

关键词： Codebook generator DNA storage Invertible neural network Self-correction

来源：评论

学校读者我要写书评

暂无评论

Exploring the Reform of Mechanical Drawing Teaching and Undergraduate Training Based on the 'Integration of Competition-Oriented Education and Industry-University Collaboration' 13

Exploring the Reform of Mechanical Drawing Teaching and Unde...

引用

13th International Conference on Educational and Information Technology, ICEIT 2024

作者： Runfeng, Zhang Yankun, Ming Zhenzhong, Liu Jianguo, Wu Xiaofei, Wang Yanfang, Sun School of Mechanical Engineering Tianjin University of Technology Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control Tianjin300384 China Tianjin University of Technology National Demonstration Center for Experimental Mechanical and Electrical Engineering Education China Tianjin Hanhailanfan Marine Technology Co. Ltd. Tianjin300457 China

ISBN: (纸本)9798350372663

This study aims to explore the reform of mechanical drawing teaching and the undergraduate training model with the goal of cultivating high-quality mechanical engineering undergraduate students. Grounded in the concepts of 'synergy between science and education' and 'integration of competition-oriented education and industry-university collaboration,' this research places talent cultivation as the central focus, linking it with discipline construction and emphasizing scientific research. The study promotes the reform of a collaborative education mechanism known as 'integration of competition-oriented education and industry-university collaboration,' effectively establishing a multi-level and multi-dimensional undergraduate training model for mechanical engineering students that incorporates the 'dual mentor system' and integrates theory, practice, and application. To apply this model in teaching practice, this research has undertaken the reform of traditional engineering drawing education, specifically addressing the issue of the gap between theory and practice in conventional teaching approaches. The proposed training program not only facilitates the development of innovative mechanical engineering professionals but also offers insights for the reform of undergraduate teaching models in the context of constructing 'new engineering disciplines'. © 2024 IEEE.

关键词： Competition

来源：评论

学校读者我要写书评

暂无评论

Development of intelligent Post-processing Algorithm for Three Degrees of Rotation Freedom CNC Machine Tools

Development of Intelligent Post-processing Algorithm for Thr...

引用

International Conference on Mechanical design, ICMD 2021

作者： Zhao, Pengrui Liu, Zhifeng Li, Zhixiong Cao, Zirui Su, Lihang Institute of Advanced Manufacturing and Intelligent Technology Beijing University of Technology Beijing100124 China Beijing Key Laboratory of Advanced Manufacturing Technology Beijing University of Technology Beijing100124 China Mechanical Industry Key Laboratory of Heavy Machine Tool Digital Design and Testing Beijing University of Technology Beijing100124 China

ISBN: (纸本)9789811673801

In this paper, the postprocessing algorithm of multi-axis CNC machine tool with three rotating degrees of freedom is studied and simulated. Firstly, the coordinate system and motion characteristics of multi-axis NC machine tool with three rotating degrees of freedom are analyzed, and the calculation method and formula of coordinate transformation of machine tool are presented. Secondly, by using MATLAB software, a universal postprocessor for NC program of three degrees of freedom machine tool is developed. Finally, the complex propeller as the processing object was used to verify the correctness of the post-processing algorithm through the VERICUT CNC machining simulation system. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Machine tools

来源：评论

学校读者我要写书评

暂无评论

Synthesis of Decoherence-Free Modes in Linear Quantum Passive Systems via Robust Pole Placement

Synthesis of Decoherence-Free Modes in Linear Quantum Passiv...

引用

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Miao, Zibo Pan, Yu Gao, Qing School of Mechanical Engineering and Automation Adaptive Robotics Harbin Institute of Technology Guangdong Key Laboratory of Intelligent Morphing Mechanisms Shenzhen518055 China The Institute of Cyber-Systems and Control College of Control Science and Engineering Zhejiang University State Key Laboratory of Industrial Control Technology Hangzhou310027 China The School of Automation Science and Electrical Engineering Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Beijing100191 China

ISBN: (纸本)9781665410205

In this paper we extend our previous research on coherent observer-based pole placement approach to study the synthesis of robust decoherence-free (DF) modes for linear quantum passive systems, which is aimed at preservation of quantum information. In particular, DF modes can be generated by placing the poles on the imaginary axis via a coherent feedback design scheme, and these modes can further be simultaneously made robust against perturbations to the system parameters by minimizing the condition number associated with imaginary poles. We develop explicit algebraic conditions for the existence of such a coherent quantum controller, with the corresponding deign procedure provided. Examples are given to illustrate the process of tuning the DF modes towards perfect robustness via the proposed pole placement technique. © 2024 IEEE.

关键词： Poles

来源：评论

学校读者我要写书评

暂无评论

An Encoded Double Auxiliary Pilot Channel Estimation Method for Intrinsic Interference Cancellation in FBMC/OQAM Systems

An Encoded Double Auxiliary Pilot Channel Estimation Method ...

引用

IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)

作者： Sixuan Xing Jianhong Xiang Ying Wang Liangang Qi Yu Zhong College of Information and Communication Engineering Harbin Engineering University Harbin China Key Laboratory of Advanced Ship Communication and Information Technology Harbin Engineering University Harbin China Agile and Intelligent Computing Key Laboratory Chengdu China

ISBN: (数字)9798350389692

ISBN: (纸本)9798350389708

Filter bank multicarrier technique (FBMC) is considered as a promising modulation technique suitable for future communications. However, the presence of intrinsic interference prevents the efficient implementation of techniques such as channel estimation. This paper presents a encode double auxiliary pilot channel estimation method (EDAP) for filter bank multicarrier offset quadrature amplitude modulation (FBMC/OQAM) systems. In order to eliminate the intrinsic interference, we add auxiliary symbols around the pilot, and at the same time perform a precoding process to minimize the auxiliary pilot energy loss and computational complexity. In order to validate the above method, the OQAM/FBMC system is modeled. Theoretical analysis and numerical results show that the method improves the channel estimation accuracy by 2.15-3.22 dB in double selective channels while maintaining low power consumption and low complexity.

关键词： Power demand Spectral efficiency Quadrature amplitude modulation System performance Precoding Channel estimation Filter banks Symbols Filtering algorithms Numerical models

来源：评论

学校读者我要写书评

暂无评论

Development and Implementation of a Wisdom Course Utilizing the "teaching-Innovation-Competition" Model within the Virtual Teaching and Research Section of Mechatronic Engineering Major in China 24

Development and Implementation of a Wisdom Course Utilizing ...

引用

3rd International Conference on Artificial Intelligence and Education, ICAIE 2024

作者： Chen, Wei Ma, Wenpeng Hu, Wenhua Liu, Fulong Zhang, Xiaotao Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control School of Mechanical Engineering Tianjin University of Technology Tianjin China Tianjin University of Technology Tianjin China

ISBN: (纸本)9798400712692

A new teaching model of teaching-innovation-competition is presented and a wisdom course based on this model is developed for mechatronics engineering majors, mainly to deal with several existing problems such as limited experimentation and lack of innovative inspiration in Chinese universities. This study elaborates the course objectives and structure of curriculum system, including digitalized teaching content, project-driven teaching means, innovation platform, web-based resource support, team competition and diversified evaluation method. Through a teaching case study within mechatronic engineering major, the teaching mechanism and achievements are demonstrated. The study results indicate that the proposed teaching mode based on the virtual teaching and research section, significantly improves students' innovation ability and the teaching effects, compared to traditional teaching methods. This teaching mode can also be extended to other majors for reference. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Teaching

来源：评论

学校读者我要写书评

暂无评论

Exploration and Practice of Cultivating Methods of Undergraduates' Ability to Apply Knowledge: Taking Engineering Drawing course as an example 23

Exploration and Practice of Cultivating Methods of Undergrad...

引用

8th International Conference on Distance Education and Learning, ICDEL 2023

作者： Wang, Xiaofei Wang, Jutao Zhang, Runfeng Yuan, Jingyao Tianjin Key Laboratory of the Design and Intelligent Control of the Advanced Mechatronical System Natl. Demonstration Ctr. for Exp. Mech. and Elec. Eng. Education Tianjin University of Technology Tianjin University of Technolog China Tianjin University of Technology China

ISBN: (纸本)9798400700422

The ability to apply knowledge is one of the most important abilities for university students. In the paper, it introduces that how to foster the ability in the teaching of engineering drawing course, and presents some views about and exploration of the reform of the engineering drawing course for cultivating engineers with modern consciousness and innovation quality. In accordance with the standards of engineering education accreditation, based on the idea of Conceive-design-Implement-Operate (CDIO), ideological and political of curriculum is merged into the course. The course reform better improve the quality of engineering education, lay a solid foundation of the specialty, and cultivate students' spatial thinking ability. At the same time, the course reform help students have a more comprehensive understanding to product-formation process, and then enhance students' sense of macro control of products. Combined with the characteristic of the engineering education accreditation and the CDIO in the teaching syllabus, curriculum structure, teaching methods and experiments are reformed. Following the pace of technological development, starting from stimulating students' interest in learning, integrating a 3D printing project into the teaching reform of engineering drawing is explored. There are two different level of introducing 3D printing into the course. They are shallow introduction and deep participation in producing a product by 3D printing. In the teaching mode of deep integration, students participate in the whole process of product design, modeling, 3D printing and all other 3D printing links. The implementation of this model is conducive to arousing students' interest in learning, giving free reign to their subjective initiative, enabling students to enjoy the learning process. Finally, the teaching objectives are realized, and their innovative ability are improved. © 2023 ACM.

关键词： Curricula

来源：评论

学校读者我要写书评

暂无评论

FLIP-80M: 80 Million Visual-Linguistic Pairs for Facial Language-Image Pre-Training 24

FLIP-80M: 80 Million Visual-Linguistic Pairs for Facial Lang...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Li, Yudong Hou, Xianxu Dezhi, Zheng Shen, Linlin Zhao, Zhe School of Computer Science and Software Engineering Shenzhen University Shenzhen China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China School of AI and Advanced Computing Xi'an Jiaotong-Liverpool University Shenzhen China Guangdong Provincial Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen China Tencent AI Lab Beijing China

ISBN: (纸本)9798400706868

While significant progress has been made in multi-modal learning driven by large-scale image-text datasets, there is still a noticeable gap in the availability of such datasets within the facial domain. To facilitate and advance the field of facial representation learning, we present FLIP-80M, a large-scale visual-linguistic dataset comprising over 80 million face images paired with text descriptions. FLIP-80M is constructed by leveraging the large openly available image-text-pair dataset LAION-5B and a mixed-method approach to filter face-related pairs from both visual and linguistic perspectives. Our curation process involves face detection, face caption classification, text de-noising, and synthesis-based image augmentation. As a result, FLIP-80M stands as the largest face-text dataset to date. To evaluate the potential of our dataset, we fine-tune the CLIP model using the proposed FLIP-80M, to create FLIP (Facial Language-Image Pretraining) and assess its representation capabilities across various downstream tasks. Our experiments demonstrate that our FLIP model achieves state-of-the-art results in a range of face analysis tasks, including face parsing, face alignment, and face attribute classification. The dataset and models are available at https://***/ydli-ai/FLIP. © 2024 ACM.

关键词： Face recognition

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：