检索结果-内蒙古大学图书馆

Selection and pruning algorithms for bitmap index selection problem using data mining

9th International Conference on Data Warehousing and Knowledge Discovery

作者： Bellatreche, Ladjel Missaoni, Rokia Necir, Hamid Drias, Habiba Univ Poitiers LISI ENSMA Poitiers France Univ Quebec Outaouais UQO Quebec City PQ Canada Inst Natl Informat Algiers Algeria

ISBN: (纸本)9783540745525

Indexing schemes are redundant structures offered by DBMSs to speed up complex queries. Two types of indices are available: mono-attribute indices (B-tree, bitmap, hash, etc.) and multi-attribute indices (join indices, bitmap join indices). In relational data warehouses, bitmap join indices (BJIs) are bitmap indices for optimizing star join queries through bit-wise operations. They can be used to avoid actual joins of tables, or to greatly reduce the volume of data that must be joined, by executing restrictions in advance. BJIs are defined using non-key dimension attributes and fact key attributes. Moreover, the problem of selecting these indices is difficult because there is a large number of candidate attributes (defined on dimension tables) that could participate in building these indices. To reduce this complexity, we propose an approach which first prunes the search space of this problem using data mining techniques, and then based on the new search space, it uses a greedy algorithm to select BJIs that minimize the cost of executing a set of queries and satisfy a storage constraint. The main peculiarity of our pruning approach, compared to the existing ones (that use only appearance frequencies of indexable attributes appearing in queries as a pruning metric), is that it uses others parameters such as the size of their dimension tables, the length of each tuple and the size of a disk page.

关键词： physical design bitmap join index pruning algorithms data mining

来源：评论

学校读者我要写书评

暂无评论

A Novel Thought of pruning algorithms: pruning Based on Less Training 16th

A Novel Thought of Pruning Algorithms: Pruning Based on Less...

引用

16th Pacific Rim International Conference on Artificial Intelligence (PRICAI)

作者： Li, Yue Zhao, Weibin Shang, Lin Nanjing Univ Dept Comp Sci & Technol Nanjing 210023 Peoples R China

ISBN: (纸本)9783030299118;9783030299101

Pre-training of models in pruning algorithms plays an important role in pruning decision-making. We find that excessive pre-training is not necessary for pruning algorithms. According to this idea, we propose a pruning thought-Incremental pruning based on less training (IPLT). We can combine IPLT with almost all existing pruning algorithms. Compared with the original pruning algorithms based on a large number of pre-training, the modified algorithms (by IPLT) has competitive compression effect. On the premise of ensuring accuracy, the pruning algorithms modified by IPLT can achieve 8x-9x compression for VGG-16 on CIFAR-10 and only needs to pre-train few epochs. For VGG-16 on CIFAR-10, we can not only achieve 10x test acceleration, but also about 10x training acceleration. At present, the research mainly focuses on the compression and acceleration in the application stage of the models, while the compression and acceleration in the training stage are few. We newly proposed the thought of IPLT that can compress and accelerate in the training stage. It is novel to consider the amount of pre-training required by pruning algorithm. Our results have implications: Too much pre-training may be not necessary for pruning algorithms.

关键词： pruning algorithms Amount of pre-training Too many

来源：评论

学校读者我要写书评

暂无评论

PoPS: Policy pruning and Shrinking for Deep Reinforcement Learning

引用

IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING 2020年第4期14卷 789-801页

作者： Livne, Dor Cohen, Kobi Ben Gurion Univ Negev Sch Elect & Comp Engn IL-8410501 Beer Sheva Israel

The recent success of deep neural networks (DNNs) for function approximation in reinforcement learning has triggered the development of Deep Reinforcement Learning (DRL) algorithms in various fields, such as robotics, computer games, natural language processing, computer vision, sensing systems, and wireless networking. Unfortunately, DNNs suffer from high computational cost and memory consumption, which limits the use of DRL algorithms in systems with limited hardware resources. In recent years, pruning algorithms have demonstrated considerable success in reducing the redundancy of DNNs in classification tasks. However, existing algorithms suffer from a significant performance reduction in the DRL domain. In this article, we develop the first effective solution to the performance reduction problem of pruning in the DRL domain, and establish a working algorithm, named Policy pruning and Shrinking (PoPS), to train DRL models with strong performance while achieving a compact representation of the DNN. The framework is based on a novel iterative policy pruning and shrinking method that leverages the power of transfer learning when training the DRL model. We present an extensive experimental study that demonstrates the strong performance of PoPS using the popular Cartpole, Lunar Lander, Pong, and Pacman environments. Finally, we develop an open source software for the benefit of researchers and developers in related fields.

关键词： Signal processing algorithms Redundancy Training Reinforcement learning Approximation algorithms Games Task analysis Deep reinforcement learning (DRL) deep neural network (DNN) pruning algorithms

来源：评论

学校读者我要写书评

暂无评论

THE VECTOR-RADIX FAST COSINE TRANSFORM - pruning AND COMPLEXITY ANALYSIS

引用

SIGNAL PROCESSING 1995年第2期43卷 197-205页

作者： CHRISTOPOULOS, CA BORMANS, J CORNELIS, J SKODRAS, AN FREE UNIV BRUSSELS ETROIRISB-1050 BRUSSELSBELGIUM

An extension of one of the fastest existing algorithms for the computation of the two-dimensional (2D) discrete cosine transform (DCT) is presented and a 2D fast cosine transform (FCT) pruning algorithm is derived. The proposed vector-radix FCT algorithm has the same computational complexity as the fastest existing 2D algorithms, but outperforms them in speed of computation, in facilitating VLSI implementation and pruning properties. For the computation of N x N DCT points, the proposed algorithm requires N-2 fewer memory locations and 4N(2) fewer data transfers. For the computation of N-0 x N-0 out of N x N points (N-0 = 2(m0)) by means of the proposed vector-radix FCT pruning algorithm, 3m(0)N(2)/4 multiplications and (2m(0) + 1)N-2 + (m(0) - 3)N-0(2) + 2N(0) additions are required, while the number of memory accesses for the computation of the butterflies is 2m(0)N(2) + 5(N-2 - N-0(2))/3. This reduced total complexity of the proposed algorithm makes it a useful tool for many image coding applications on general and special purpose computer platforms.

关键词： FAST algorithms DISCRETE COSINE TRANSFORM pruning algorithms

来源：评论

学校读者我要写书评

暂无评论

Bio-inspired machine learning: programmed death and replication

引用

NEURAL COMPUTING & APPLICATIONS 2023年第27期35卷 20273-20298页

作者： Grabovsky, Andrey Vanchurin, Vitaly Budker Inst Nucl Phys Novosibirsk 630090 Russia Novosibirsk State Univ Novosibirsk 630090 Russia Natl Ctr Biotechnol Informat NIH Bethesda MD 20894 USA Duluth Inst Adv Study Duluth MN 55804 USA

We analyze algorithmic and computational aspects of biological phenomena, such as replication and programmed death, in the context of machine learning. We use two different measures of neuron efficiency to develop machine learning algorithms for adding neurons to the system (i.e., replication algorithm) and removing neurons from the system (i.e., programmed death algorithm). We argue that the programmed death algorithm can be used for compression of neural networks and the replication algorithm can be used for improving performance of the already trained neural networks. We also show that a combined algorithm of programmed death and replication can improve the learning efficiency of arbitrary machine learning systems. The computational advantages of the bio-inspired algorithms are demonstrated by training feedforward neural networks on the MNIST dataset of handwritten images.

关键词： Machine learning Neural networks Bio-inspired algorithms Neuron correlations pruning algorithms Constructive algorithms Classification

来源：评论

学校读者我要写书评

暂无评论

Lazy evaluations in Lukasiewicz type fuzzy logic

引用

FUZZY SETS AND SYSTEMS 2019年 376卷 127-151页

作者： Nagy, Benedek Basbous, Raed Tajti, Tibor Eastern Mediterranean Univ Fac Arts & Sci Dept Math Via Mersin 10 Famagusta Northern Cyprus Turkey Al Quds Open Univ Fac Technol & Appl Sci Jerusalem Palestine Eszterhazy Karoly Univ Eger Hungary

Lazy evaluations are playing important roles in various fields of computer science including hardware design, programming, decision making. In this paper one of the most known and used fuzzy logic system, the Lukasiewicz logic is considered. pruning algorithms are presented to quicken the evaluations of logical formulae in Lukasiewicz logic by cutting out those branches of the formula tree that have no influence on the final result for some reasons. Correctness of the algorithm is also proven. (C) 2018 Elsevier B.V. All rights reserved.

关键词： Non-classical logics pruning algorithms Formula trees Fast evaluation Fuzzy logic Many valued logic Lazy evaluations

来源：评论

学校读者我要写书评

暂无评论

Optimising an observational water monitoring network for Archipelago Sea, South West Finland

引用

COMPUTERS & GEOSCIENCES 2011年第7期37卷 844-854页

作者： Murtojarvi, Mika Suominen, Tapio Uusipaikka, Esa Nevalainen, Olli S. Univ Turku Dept Informat Technol FI-20014 Turku Finland Univ Turku Turku Ctr Comp Sci TUCS FI-20014 Turku Finland Univ Turku Dept Geog FI-20014 Turku Finland Univ Turku Dept Stat FI-20014 Turku Finland

Water quality monitoring in topographically fragmented archipelago coasts calls for a dense observational network. However, visiting multiple sites and analyzing the samples requires a significant amount of work, leading to considerable economic cost. It is of interest to determine an efficient set of sites, which still offers adequate information on the water quality with a sufficient spatial accuracy. A method for optimizing an existing observational network is proposed. The method is concretized by applying it for an observational network in the Archipelago Sea, South West Finland. The network is pruned with the requirement that the observations of the removed sites can be estimated using those of the remaining sites. Suboptimal heuristics are used in pruning to keep the computational time acceptable. Some observations are not available and need to be estimated (imputed) before the pruning. For the network in the Archipelago Sea, the results of the pruning are somewhat sensitive to differences in imputed datasets and heuristics used for site selection. (C) 2011 Elsevier Ltd. All rights reserved.

关键词： Water quality The Baltic Sea Observational network pruning algorithms

来源：评论

学校读者我要写书评

暂无评论

A dynamic ensemble learning algorithm for neural networks

引用

NEURAL COMPUTING & APPLICATIONS 2020年第12期32卷 8675-8690页

作者： Alam, Kazi Md Rokibul Siddique, Nazmul Adeli, Hojjat Khulna Univ Engn & Technol Dept Comp Sci & Engn Khulna 9203 Bangladesh Ulster Univ Sch Comp Engn & Intelligent Syst Derry BT48 7JL Londonderry North Ireland Ohio State Univ Dept Neurosci Columbus OH 43210 USA Ohio State Univ Dept Neurol Columbus OH 43210 USA Ohio State Univ Dept Biomed Informat Columbus OH 43210 USA

This paper presents a novel dynamic ensemble learning (DEL) algorithm for designing ensemble of neural networks (NNs). DEL algorithm determines the size of ensemble, the number of individual NNs employing a constructive strategy, the number of hidden nodes of individual NNs employing a constructive-pruning strategy, and different training samples for individual NN's learning. For diversity, negative correlation learning has been introduced and also variation of training samples has been made for individual NNs that provide better learning from the whole training samples. The major benefits of the proposed DEL compared to existing ensemble algorithms are (1) automatic design of ensemble;(2) maintaining accuracy and diversity of NNs at the same time;and (3) minimum number of parameters to be defined by user. DEL algorithm is applied to a set of real-world classification problems such as the cancer, diabetes, heart disease, thyroid, credit card, glass, gene, horse, letter recognition, mushroom, and soybean datasets. It has been confirmed by experimental results that DEL produces dynamic NN ensembles of appropriate architecture and diversity that demonstrate good generalization ability.

关键词： Neural network ensemble Backpropagation algorithm Negative correlation learning Constructive algorithms pruning algorithms

来源：评论

学校读者我要写书评

暂无评论

Approximating the sheep milk production curve through the use of artificial neural networks and genetic algorithms

引用

COMPUTERS & OPERATIONS RESEARCH 2005年第10期32卷 2653-2670页

作者： Torres, M Hervás, C Amador, F Univ Cordoba ETEA Fac Ciencias Econ & Empresariales E-14004 Cordoba Spain

This paper examines the potential of a neural network coupled with genetic algorithms to recognize the parameters that define the production curve of sheep milk, in which production is time-dependent, using solely the data registered in the animals' first controls. This enables the productive capacity of the animal to be identified more rapidly and leads to a faster selection process in determining the best producers. For this purpose we employ a network with a single hidden layer, using the property of "universal approximation". To find the number of nodes to be included in this layer, genetic and pruning algorithms are applied. Results thus obtained applying genetic and pruning algorithms are found to be better than other models which exclusively apply the classical learning algorithm Extended-Delta-Bar-Delta. (c) 2004 Elsevier Ltd. All rights reserved.

关键词： artificial neural networks genetic algorithms pruning algorithms non-linear regression extended-Delta-Bar-Delta gamma function

来源：评论

学校读者我要写书评

暂无评论

Querying Optimal Routes for Group Meetup

引用

DATA SCIENCE AND ENGINEERING 2021年第2期6卷 180-191页

作者： Chen, Bo Zhu, Huaijie Liu, Wei Yin, Jian Lee, Wang-Chien Xu, Jianliang Sun Yat Sen Univ Sch Data & Comp Sci Guangzhou Peoples R China Penn State Univ Dept Comp Sci & Engn State Coll PA USA Hong Kong Baptist Univ Kowloon Tong Hong Kong Peoples R China

Motivated by location-based social networks which allow people to access location-based services as a group, we study a novel variant of optimal sequenced route (OSR) queries, optimal sequenced route for group meetup (OSR-G) queries. OSR-G query aims to find the optimal meeting POI (point of interest) such that the maximum users' route distance to the meeting POI is minimized after each user visits a number of POIs of specific categories (e.g., gas stations, restaurants, and shopping malls) in a particular order. To process OSR-G queries, we first propose an OSR-Based (OSRB) algorithm as our baseline, which examines every POI in the meeting category and utilizes existing OSR (called E-OSR) algorithm to compute the optimal route for each user to the meeting POI. To address the shortcomings (i.e., requiring to examine every POI in the meeting category) of OSRB, we propose an upper bound based filtering algorithm, called circle filtering (CF) algorithm, which exploits the circle property to filter the unpromising meeting POIs. In addition, we propose a lower bound based pruning (LBP) algorithm, namely LBP-SP which exploits a shortest path lower bound to prune the unqualified meeting POIs to reduce the search space. Furthermore, we develop an approximate algorithm, namely APS, to accelerate OSR-G queries with a good approximation ratio. Finally the experimental results show that both CF and LBP-SP outperform the OSRB algorithm and have high pruning rates. Moreover, the proposed approximate algorithm runs faster than the exact OSR-G algorithms and has a good approximation ratio.

关键词： Route queries Group meeting pruning algorithms

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：