检索结果-内蒙古大学图书馆

Automatics and Informatics (ICAI), International Conference

作者： Palvan I. Kalandarov Shakhnoza R. Ubaydullayeva Mirhalil A. Ismailov Nikola N. Nikolov Mariela Alexandrova Department "Automation and Control of Technology Process in Production" «Tashkent Institute of Irrigation and Agricultural Mechanization Engineers» National Research University Tashkent Uzbekistan Department of Automation Faculty of Computer Science and Automation Technical University of Varna Varna Bulgaria

ISBN: (数字)9798350353907

ISBN: (纸本)9798350353914

The article discusses the theoretical foundations of the design of a single-channel ultrahigh frequency moisture meter with direct measurement of the moisture content of bulk materials. In accordance with the requirements of the methodology for selecting efficiency criteria, it is necessary to develop structural diagrams of measuring devices. For these purposes, the standard deviation of the random error is determined, characterizing the accuracy. It includes the main components: sensitivity error, zero error and additive component. Mathematical models of structures are constructed and the standard deviation of random errors, which are caused by certain parameters and additive fluctuations, is calculated. A single-parameter ultrahigh-frequency method for determining the moisture content is proposed. This method provides high accuracy of a single-channel ultrahigh frequency moisture meter with direct measurement of the moisture content of bulk materials. The measuring device can be used in the agricultural industry, where humidity is one of the important parameters, starting with harvesting and ending with the release of finished products.

关键词： Meters Additives Accuracy Sensitivity Moisture UHF measurements Mathematical models Frequency measurement Informatics Standards

来源：评论

学校读者我要写书评

暂无评论

Optimizing Attention for Efficient LLM Inference: A Review

Optimizing Attention for Efficient LLM Inference: A Review

引用

World Congress on Computing and Communication Technologies (WCCCT)

作者： Siyuan Sun Jinling Yu Hanshuo Liu Hanyun Guo Yang Cao Shouhua Zhang Jiehan Zhou School of Computer Science and Engineering Shandong University of Science and Technology Qingdao China Faculty of Computer Science Western University London Canada Research&Development Department Mox Automation Shanghai Company Shanghai China Faculty of Information Technology and Electrical Engineering University of Oulu Oulu Finland

ISBN: (数字)9798331512620

ISBN: (纸本)9798331512637

The rapid advancement of deep learning has led to significant progress in large language models (LLMs), with the Attention mechanism serving as a core component of their success. However, the computational and memory demands of Attention mechanisms pose bottlenecks for efficient inference, especially in long-sequence and real-time tasks. This paper systematically reviews optimization strategies for Attention mechanisms, including sparse attention, low-rank decomposition, quantization techniques, block-based parallel computation, and memory management. These approaches have demonstrated notable improvements in reducing computational complexity, optimizing memory usage, and enhancing inference performance. This review highlights the key challenges of computational efficiency, long-sequence modeling, and cross-task generalization through an in-depth analysis of existing methods, their advantages, and limitations. Future research directions, including dynamic precision, hardware-aware optimization, and lightweight architectures offer insights for advancing LLM inference theory and practice.

关键词： Attention mechanisms Quantization (signal) Reviews Computational modeling Large language models Memory management Optimization methods Real-time systems Computational efficiency Computational complexity

来源：评论

学校读者我要写书评

暂无评论

Sports Programming as a Tool for Regional Development of the Digital Industry and Higher Education 4

Sports Programming as a Tool for Regional Development of the...

引用

4th International Conference on Technology Enhanced Learning in Higher Education, TELE 2024

作者： Galkin, Alexander Miroshnikov, Artem Sapronov, Daniil Skarzhinskaya, Elena Lipetsk State Technical University Faculty of Automation and Computer Science Lipetsk Russia Lipetsk State Technical University Department of Applied Mathematics Lipetsk Russia The Russian Federation of Sports Programming Lipetsk branch Lipetsk Russia Moscow Financial and Industrial University 'Synergy' Department of Cybersport Moscow Russia

ISBN: (纸本)9798350353532

The article presents the history of the development of sports programming as a sport in Russia. The main stages of the formation of the legislative basis for this sport are given. The rationale and description of the process of training athletes in higher education institutions are given - possible traj ectories in the disciplines of 'algorithmic programming' and 'product programming', the influence of the development of sports programming on improving the quality of IT education. As an example, the organization of training students in the discipline of 'algorithmic programming' at the Lipetsk State Technical University is given. The activities of the Lipetsk regional branch of the All- Russian Sports Federation to popularize this sport including among students are also described. © 2024 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

SGTC: Semantic-Guided Triplet Co-training for Sparsely Annotated Semi-Supervised Medical Image Segmentation

arXiv

引用

arXiv 2024年

作者： Yan, Ke Cai, Qing Zhang, Fan Cao, Ziyan Liu, Zhi Faculty of Computer Science and Technology Ocean University of China China School of Automation Northwestern Polytechnical University China School of Information Science and Engineering Shandong University China

Although semi-supervised learning has made significant advances in the field of medical image segmentation, fully annotating a volumetric sample slice by slice remains a costly and time-consuming task. Even worse, most of the existing approaches pay much attention to image-level information and ignore semantic features, resulting in the inability to perceive weak boundaries. To address these issues, we propose a novel Semantic-Guided Triplet Co-training (SGTC) framework, which achieves high-end medical image segmentation by only annotating three orthogonal slices of a few volumetric samples, significantly alleviating the burden of radiologists. Our method consist of two main components. Specifically, to enable semantic-aware, fine-granular segmentation and enhance the quality of pseudo-labels, a novel semantic-guided auxiliary learning mechanism is proposed based on the pretrained CLIP. In addition, focusing on a more challenging but clinically realistic scenario, a new triple-view disparity training strategy is proposed, which uses sparse annotations (i.e., only three labeled slices of a few volumes) to perform co-training between three sub-networks, significantly improving the robustness. Extensive experiments on three public medical datasets demonstrate that our method outperforms most state-of-the-art semi-supervised counterparts under sparse annotation settings. The source code is available at https://***/xmeimeimei/SGTC. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Accurate Prediction of Workloads and Resources with Multi-Head Attention and Hybrid LSTM for Cloud Data Centers

Accurate Prediction of Workloads and Resources with Multi-He...

引用

作者： Bi, Jing Ma, Haisen Yuan, Haitao Zhang, Jia Beijing University of Technology School of Software Engineering Faculty of Information Technology Beijing100124 China Beihang University School of Automation Science and Electrical Engineering Beijing100191 China Southern Methodist University Department of Computer Science Lyle School of Engineering DallasTX75205 United States

Currently, cloud computing service providers face big challenges in predicting large-scale workload and resource usage time series. Due to the difficulty in capturing nonlinear features, traditional forecasting methods usually fail to achieve high prediction performance for resource usage and workload sequences. Besides, there is much noise in original time series of resources and workloads. If these time series are not de-noised by smoothing algorithms, the prediction results can fail to meet the providers' requirements. To do so, this work proposes a hybrid prediction model named VAMBiG that integrates Variational mode decomposition, an Adaptive Savitzky-Golay (SG) filter, a Multi-head attention mechanism, Bidirectional and Grid versions of Long and Short Term Memory (LSTM) networks. VAMBiG adopts a signal decomposition method named variational mode decomposition to decompose complex and non-linear original time series into low-frequency intrinsic mode functions. Then, it adopts an adaptive SG filter as a data pre-processing tool to eliminate noise and extreme points in such functions. Afterwards, it adopts bidirectional and grid LSTM networks to capture bidirectional features and dimension ones, respectively. Finally, it adopts a multi-head attention mechanism to explore importance of different data dimensions. VAMBiG aims to predict resource usage and workloads in highly variable traces in clouds. Extensive experimental results demonstrate that it achieves higher-Accuracy prediction than several advanced prediction approaches with datasets from Google and Alibaba cluster traces. © 2016 IEEE.

关键词： Time series analysis

来源：评论

学校读者我要写书评

暂无评论

Improved Adaptive Particle Swarm Optimization Algorithm with a Two-Way Learning Method 4th

Improved Adaptive Particle Swarm Optimization Algorithm with...

引用

4th International Conference on Wireless Communications and Applications, ICWCA 2020

作者： Wang, Yawen Qian, Qian Feng, Yong Fu, Yunfa Faculty of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming650500 China Yunnan Key Laboratory of Computer Technology Applications Kunming University of Science and Technology Yunnan Kunming650500 China

ISBN: (纸本)9789811651632

Traditional particle swarm optimization algorithm has some disadvantages, such as slow convergence speed and easy to fall into local extremes. In order to improve the performance, an improved adaptive particle swarm optimization algorithm with a two-way learning method is proposed. First, the algorithm adaptively adjusts the algorithm according to the iteration periods of the optimization process. Specifically, the inertia weight and the value of the learning factor are changed nonlinearly, so as to better balance the search behavior of the particles in the group;Second, the idea of beetle search is introduced into the particle swarm algorithm to form a new two-way learning mechanism, which overcomes the limitations of the traditional particle swarm algorithm and help to increase the diversity of the population. In this way, the search scope is expanded, and the search accuracy of the algorithm is enhanced. Finally, the simulation is carried out on several multi-dimensional functions, and compared with other two related algorithms. The experimental results show that under the same experimental conditions, the improved algorithm has obvious advantages in optimization ability and convergence speed. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Particle swarm optimization (PSO)

来源：评论

学校读者我要写书评

暂无评论

A loading path planning method for co-loading in automatic harvesting system

A loading path planning method for co-loading in automatic h...

引用

2023 International Conference on Advanced Mechatronic Systems, ICAMechS 2023

作者： Wang, Xiaoxiong Wang, Zhuo Bai, Xiaoping Ge, Zhikang Wang, Jian Man, Zhihong Chinese Academy of Sciences Institutes for Robotics and Intelligent Manufacturing Shenyang China University of Chinese Academy of Sciences School of Computer Science and Technology Beijing China Chinese Academy of Sciences Shenyang Institute of Automation Shenyang China Swinburne University of Technology Faculty of Science Engineering and Technology Melbourne Australia

ISBN: (纸本)9798350307474

In this paper, a path planning method is developed for co-loading in master-slave automatic harvesting. It is shown that, the adjustment of the unloading port position relative to the grain tank can be performed jointly by the longitudinal relative position between the vehicles and the angle of the unloading arm. Then the loading path based on the Hamiltonian circuit is designed to guide the motion of the unloading port during the loading process. By improving the loading uniformity, the loading path proposed can maximize the capacity utilization of grain tank which can reduce the round-trip frequency of the transport vehicle and improve the energy and time efficiency of the harvesting operation. Simulation results are provided to validate the excellent performance of presented path planning method. © 2023 IEEE.

关键词： Motion planning

来源：评论

学校读者我要写书评

暂无评论

Sparrow search algorithm based on cosine similarity and random multi-chaotic perturbation 6

Sparrow search algorithm based on cosine similarity and rand...

引用

6th International Conference on Mechatronics and Intelligent Robotics, ICMIR 2022

作者： Li, Guangyang Qian, Qian Feng, Yong Fu, Yunfa Faculty of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming650500 China Yunnan Key Laboratory of Computer Technology Applications Kunming University of Science and Technology Yunnan Kunming650500 China

ISBN: (纸本)9781510656635

Aiming at the problems of sparrow algorithm (SSA), such as easy to fall into local extremum, uneven initial population distribution and slow convergence in late iteration, a sparrow search algorithm (CMSSA) was proposed, which combined cosine similarity and random multi-chaotic disturbance. The algorithm firstly integrates the reverse learning strategy and initializes the population by using the population uniform adjustment strategy of cosine similarity to ensure the uniformity and richness of the population, so that the algorithm can better search for the global optimal solution. Secondly, a random selection mechanism of multi-chaos local search strategy is used to take advantages from different disturbance states of multiple chaos models. Randomly selected chaos maps are used in each iteration to perturb individuals and help SSA get rid of local extremums. Simulation results show that compared with other intelligent algorithms,CMSSA achieves better results in robustness and optimization accuracy. © 2022 SPIE.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Flower Pollination Algorithm Based on Beetle Antennae Search Method 4th

Flower Pollination Algorithm Based on Beetle Antennae Search...

引用

4th International Conference on Wireless Communications and Applications, ICWCA 2020

作者： Zhou, Jiayi Qian, Qian Fu, Yunfa Feng, Yong Faculty of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming650500 China Yunnan Key Laboratory of Computer Technology Applications Kunming University of Science and Technology Yunnan Kunming650500 China

ISBN: (纸本)9789811651632

In this paper, a new flower pollination algorithm based on beetle antennae search method (TBFPA) is proposed to deal with the slow convergence speed problem of traditional flower pollination algorithm. Specifically, after the new individuals are generated by the way of flower pollination algorithm, the beetle antennae search method is used to search the nearby solution space of the new individuals, thus increasing the probability to reach better solutions and speeding up the convergence speed of the algorithm. In addition, a new uniform distribution sampling method is used to generate the initial population so as to spread the initial solutions all over the solution space. In order to verify the effectiveness of TBFPA, different standard functions are tested between TBFPA and FPA algorithms and the results proved that TBFPA has faster convergence speed and stronger global searchability. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Beetle antennae search algorithm Flower pollination Function optimization

来源：评论

学校读者我要写书评

暂无评论

A Robust Adaptive Terminal Sliding Mode Control for Lane-keeping in Road Vehicles

A Robust Adaptive Terminal Sliding Mode Control for Lane-kee...

引用

2023 International Conference on Advanced Mechatronic Systems, ICAMechS 2023

作者： Ge, Zhikang Wang, Zhuo Bai, Xiaoping Wang, Xiaoxiong Man, Zhihong Chinese Academy of Sciences Institutes for Robotics and Intelligent Manufacturing Shenyang China University of Chinese Academy of Sciences School of Computer Science and Technology Beijing China Chinese Academy of Sciences Shenyang Institute of Automation Shenyang China Swinburne University of Technology Faculty of Science Engineering and Technology Melbourne Australia

ISBN: (纸本)9798350307474

In this paper, a novel adaptive terminal sliding-mode control (ATSMC) method is developed for road vehicles with uncertain dynamics. It is shown that the designed adaptive laws can recursively update the controller parameters based on the estimation values of the unknown parameter bound in the Lyapunov sense, and then the lateral position error can be driven to zero in a finite time. In comparison to the conventional sliding-mode-based vehicle lateral control method, the proposed ATSMC not only ensures fast error convergence, but also eliminates the requirement for known bounded information in controller design. Simulation results are presented to demonstrate the good performance of the proposed ATSMC scheme. © 2023 IEEE.

关键词： Sliding mode control

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：