检索结果-内蒙古大学图书馆

International Conference on Management, Manufacturing and Materials Engineering (ICMMM 2011)

作者： Wang, Lin Ma, Jinwen Peking Univ Dept Informat Sci Sch Math Sci Beijing 100871 Peoples R China

ISBN: (纸本)9783037853511

For the Gaussian mixture learning, the expectation-maximization (EM) algorithm as well as its modified versions are widely used, but there are still two major limitations: (i). the number of components or Gaussians must be known in advance, and (ii). There is no generally accepted method for parameters initialization to prevent the algorithm being trapped in one of the local maxima of the likelihood function. In order to overcome these weaknesses, we proposed a greedy EM algorithm based on a kurtosis and skewness criterion. Specifically, we start with a single component and add one component step by step under the framework of EM algorithm in order to decrease the value of the kurtosis and skewness measure which provides an efficient index to show how well the Gaussian mixture model fits the sample data. In such a way, the number of components can be selected adaptively during the EM learning and the learning parameters can possibly escape from local maxima.

关键词： EM algorithm Gaussian mixture model selection greedy algorithm kurtosis and skewness criterion

来源：评论

学校读者我要写书评

暂无评论

greedy Search algorithm for Mixed Precision in Post-Training Quantization of Convolutional Neural Network Inspired by Submodular Optimization 13

Greedy Search Algorithm for Mixed Precision in Post-Training...

引用

13th Asian Conference on Machine Learning (ACML)

作者： Tsuji, Satoki Kawaguchi, Hiroshi Inoue, Atsuki Sakai, Yasufumi Yamada, Fuyuka Fujitsu Ltd Fujitsu Res Nakahara Ward 4-1-1 Kamikodanaka Kawasaki Kanagawa 2118588 Japan Kobe Univ Grad Sch Sci Technol & Innovat Nada Ward 1-1 Rokkoudai Kobe Hyogo 6578501 Japan

For lower bit-widths such as less than 8-bit, many quantization strategies include re-training in order to recover accuracy degradation. However, the re-training works against rapid deployment for wide distribution of quantized models. Therefore, post-training quantization has been getting more attention in recent years. In one example, partial quantization according to the layer sensitivity based on the accuracy after each quantization has been proposed;however, the effects of one layer quantization on the other layers has not taken into account. To further reduce the accuracy degradation, we propose a quantization scheme that considers the effects by continuously updating the accuracy after each layer quantization. Additionally, for more data compression, we extend that scheme to mixed precision, which applies a layer-by-layer fitted bit-width. Since the search space for bit allocation per layer increases exponentially with the number of layers N, existing methods require computationally intensive approach such as network training. Here, we derive practical solutions to the bit allocation problem in polynomial time O(N-2) using a deterministic greedy search algorithm inspired by submodular optimization without any training. For example, the proposed algorithm completes a search on ResNet18 for ImageNet in 1 hour for a single GPU. Compared to the case without updating the layer sensitivity, our method improves the accuracy of the quantized model by more than 1% with multiple convolutional neural networks. For examples, 6-bit quantization of MobileNetV2 achieves 80.1% reduction of model size with -1.10% accuracy degradation. 4-bit quantization of ResNet50 achieves 82.9% size reduction with -0.194% accuracy degradation. Furthermore, results show that the proposed method reduces the accuracy degradation by more than about 0.7% compared to various latest post-training quantization strategies.

关键词： Deep learning greedy algorithm Neural network Quantization Submodular optimization

来源：评论

学校读者我要写书评

暂无评论

A Graph Split on greedy-algorithm-based Self Organization Protocols

A Graph Split on Greedy-algorithm-based Self Organization Pr...

引用

3rd International Conference on Information Security and Intelligent Control (ISIC)

作者： Yin, Hu Lv, Yunfei Wuhan Second Ship Design Inst Wuhan Hubei Province Peoples R China

ISBN: (纸本)9781467325882

Recently, Self organization protocol has gained great focus from the research area for its efficient to save energy on wireless sensors network. We use a random graph model to compare some of the proposed self organization protocols in this thesis. Results show the impact of the surface on the percentage of the active set Thus, we focus on the graph splitting and the impact for the active set and the topology connectivity. Finally, we analyze the performance resulting for transmitted and received packets.

关键词： Graph split greedy algorithm self organization protocol wireless sensors network

来源：评论

学校读者我要写书评

暂无评论

A greedy GENETIC algorithm FOR UNCONSTRAINED GLOBAL OPTIMIZATION

引用

Journal of Systems Science & Complexity 2005年第1期18卷 102-110页

作者： ZHAOXinchao KeyLaboratoryofMathematicsMechanization InstituteofSystemsScienceAMSSChineseAcademyofSciencesBeijing100080China

The greedy algorithm is a strong local searching algorithm. The genetic algorithm is generally applied to the global optimization problems. In this paper, we combine the greedy idea and the genetic algorithm to propose the greedy genetic algorithm which incorporates the global exploring ability of the genetic algorithm and the local convergent ability of the greedy algorithm. Experimental results show that greedy genetic algorithm gives much better results than the classical genetic algorithm.

关键词： genetic algorithm greedy algorithm greedy genetic algorithm globaloptimization

来源：评论

学校读者我要写书评

暂无评论

A Communication Contention-Aware greedy algorithm for Scheduling Join Task Graphs

A Communication Contention-Aware Greedy Algorithm for Schedu...

引用

World Automation Congress

作者： Jianjun Zhang Dengbin Huang Yong Qu College of Science Naval University of Engineering 430033 Wuhan Hubei Province China

ISBN: (纸本)9781467344975

Aiming at the problem that many previous scheduling algorithms for task graphs always ignore the communication contention, this paper presents a communication contention-aware greedy scheduling algorithm, named CCAG_J, for scheduling join task graphs, which can improve the scheduling performance by serializing the communication edges to integrate the communication awareness into task scheduling. Experimental evaluation validated that the CCAG_J algorithm produces superior results than other compared algorithms in terms of schedule length, number of used processors, speedup and efficiency.

关键词： Task scheduling Join task graph greedy algorithm Communication contention Critical task

来源：评论

学校读者我要写书评

暂无评论

RADIO RESOURCE ALLOCATION BASED ON greedy algorithm AND SUCCESSIVE INTERFERENCE CANCELLATION IN DEVICE-TO-DEVICE (D2D) COMMUNICATION

RADIO RESOURCE ALLOCATION BASED ON GREEDY ALGORITHM AND SUCC...

引用

IET International Conference on Information and Communications Technologies

作者： Yu Tao Jun Sun Shixiang Shao Nanjing University of Posts and Telecommunications Jiangsu Provincial Key Laboratory of Wireless Communication 210003 China|c|

ISBN: (纸本)9781629931357

Device-to-Device (D2D) communication underlaying cellular networks can improve resource utilization and thus increase total throughput. However, resource allocation and interference coordination between cellular networks and D2D system will become critical and need to be properly handled. According to some papers, which just allowed a set of cellular user equipment (UE) and a pair of D2D UEs to coexist in a cellular channel, the growth of resource utilization is not sufficient. In this paper, we propose a method of radio resource allocation based on greedy algorithm and successive interference cancellation (SIC), which allows the coexistence of a cellular UE and three pairs of D2D UEs in a channel. The greedy algorithm is applied to maximize the performance of D2D links and then the SIC technique is used to cope with the interference from D2D UEs to cellular UEs. Simulation results show that the proposed method can achieve substantial gains in terms of resource utilization and total cell throughput.

关键词： Device-to-Device (D2D) communication Radio resource allocation Interference coordination greedy algorithm Successive interference cancellation (SIC)

来源：评论

学校读者我要写书评

暂无评论

Optimal Placement of PMUs using greedy algorithm and State Estimation

Optimal Placement of PMUs using Greedy Algorithm and State E...

引用

International Conference on Power Electronics, Intelligent Control and Energy Systems

作者： Varsha Jaiswal Siddhartha Sankar Thakur Biswaranjan Mishra School of Electrical and Electronics Lovely Professional University Electrical Engineering Department National Institute of Technology

ISBN: (纸本)9781467385886

This paper proposes a new technique of state estimation (SE) for electric power systems. In the proposed scheme, the Phasor Measurement Units (PMU) are first placed optimally using greedy algorithm for cost reduction, while complete observability of system is also obtained. The SE uses a linear measurement model to obtain the estimated states directly, without any iteration, thereby improves the quality of the estimated data base. To reveal the efficacy of the proposed scheme it has been tested on standard IEEE 5-bus, I4-bus, 30-bus, 57-bus and 118-Bus test systems and the test results are presented.

关键词： Phasor Measurement Unit Synchronized Measurement Optimal PMU Placement greedy algorithm

来源：评论

学校读者我要写书评

暂无评论

Segmentation of Leaf Images using greedy algorithm

Segmentation of Leaf Images using Greedy Algorithm

引用

International Conference on Energy, Communication, Data Analytics and Soft Computing

作者： Bhagya M. Patil Basavaraj Amarapur Department of Computer Science & Engineering PES University Bangalore India Department of Electrical and Electronics Engineering PDA College of Engineering Gulbarga India

ISBN: (纸本)9781538618882

The Leaf is the important part of the plant which contains important information which will be playing a role in the identification and classification of plants. The identification of plants can be done with the help of image processing techniques. The techniques are used to understand, to analyze, to interpret and to get better quality of the images for human/machine perception. The image processing includes image pre-processing, segmentation, feature extraction and classification. In this paper, the leaf segmentation of different plant such as Jackfruit, Banaba, Cotton and etc. have been experimented using greedy snake algorithm and it is compared with the M Kass snake algorithm. From the comparison, it is observed that greedy algorithm is faster and efficient than the Kass algorithm in terms of iterations obligatory to get the desired contour of an image.

关键词： Image processing greedy algorithm Kass snake algorithm Leaf segmentation

来源：评论

学校读者我要写书评

暂无评论

Selecting Sequence of Refactoring Techniques Usage for Code Changing Using greedy algorithm

Selecting Sequence of Refactoring Techniques Usage for Code ...

引用

2013 IEEE 4th International Conference on Electronics Information and Emergency Communication

作者： Ratapong Wongpiang Pornsiri Muenchaisri Center of Excellence in Software Engineering Department of Computer Engineering Chulalongkorn University Bangkok 10330Thailand

ISBN: (纸本)9781467349345

Refactoring is the process of changing the internal structure of software but it preserves external behavior of software. To improve software maintainability, we may apply several refactoring techniques to source code. Applying different sequence of refactoring techniques to different parts of source code results in different code changes and different level of software maintainability. We propose an approach for selecting sequence of refactoring techniques usage for code changing using greedy algorithm. To get optimal software maintainability, we create possible sequences of refactoring techniques usage and apply each refactoring technique to source code. greedy algorithm is used to separate the optimal sequence of refactoring techniques usage from possible sequences of refactoring techniques. We evaluate the approach with source code containing Long Method, Large Class and Feature Envy bad smells by comparing the changed source code result between applying the approach and without applying sequence of refactoring techniques usage. The compared results show that the changed source code by applying our approach can improve software maintainability better than the changed source code without sequencing refactoring techniques usage.

关键词： Refactoring Refactoring Sequencing Software Maintainability greedy algorithm

来源：评论

学校读者我要写书评

暂无评论

Modeling and evaluating the travel behaviour in multimodal networks: A path-based unified equilibrium model and a tailored greedy solution algorithm

引用

TRANSPORTATION RESEARCH PART A-POLICY AND PRACTICE 2024年 182卷

作者： Fan, Yinchao Ding, Jianxun Long, Jiancheng Wu, Jianjun Hefei Univ Technol Sch Management Hefei 230009 Peoples R China Hefei Univ Technol Sch Automot & Transportat Engn Hefei 230009 Peoples R China Engn Res Ctr Intelligent Transportat & Cooperat Ve Hefei 230009 Peoples R China Beijing Jiaotong Univ State Key Lab Rail Traff Control & Safety Beijing 100044 Peoples R China

The modeling and efficient solution of the combined mode split and traffic assignment (CMSTA) problem serve as a powerful tool for capturing complex travel behavior in multimodal transportation networks under different planning scenarios and incentive programs. In this paper, we propose a path -based unified equilibrium condition that combines the cross -nested logit (CNL)based mode split and the user equilibrium (UE)-based traffic assignment to address the CMSTA problem on a multimodal transportation network. The equilibrium condition is further formulated as a novel path -based variational inequality (VI) model. A general path -based algorithm framework that integrates a tailored greedy algorithm and a novel modified intelligent acceleration strategy (MIAS) is then developed for solving the proposed CMSTA model. Numerical examples demonstrate the effectiveness and efficiency of the proposed model and algorithm in both small -size and large-scale networks. The proposed model and algorithm can help to provide some policy implications for multimodal transportation planning and management. The lessons learned from our analysis results include (1) the removal of some existing key park -and -ride (P &R) interchanges that carry more flows from the multimodal network can result in a significant increase in total travel costs, but the removal of others may instead reduce total travel costs;(2) increasing the number of P &R interchanges in a multimodal network may degrade the performance of the network, even though it may encourage the use of green travel modes.

关键词： Multimodal transportation network Combined mode split and traffic assignment greedy algorithm Policy implication

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：