检索结果-内蒙古大学图书馆

IEEE Conference on Wireless Communications and Networking

作者： Robert Rauch Juraj Gazda Marco Levorato Department of Computers and Informatics Technical University of Kosice Kosice Slovakia Computer Science Department University of California Irvine Irvine USA

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

In response to the escalating demand for high-performance applications in Connected Autonomous Vehicles (CAVs), there has been an increased effort to accelerate Computer Vision (CV) tasks. Innovations in the field suggest that partitioning Deep Learning (D L) models, which are pivotal for CV tasks, can effectively utilize the computational strengths of both CAVs and proximal edge servers. This is achieved by splitting the DL models into two parts, where execution begins on the CAV, then the intermediate results are offloaded to complete processing on the edge server. Moreover, the precision and response time of these applications vary significantly, necessitating the adoption of early exit strategies. These strategies involve modifying DL models with auxiliary branches that allow for early termination at various stages, enabling the model to exit processing early when application requirements are met. To navigate the trade-offs between accuracy and latency, it is imperative to choose the model's optimal split and exit point that align with the current environmental conditions, such as server data throughput and computational load. To achieve this, we employ a state-of-the-art method that utilizes the Deep Deterministic Policy Gradient (DDPG) algorithm, augmented by Long Short-Term Memory networks. Additionally, to refine the training of the DDPG's actor network, Imitation Learning (IL) is integrated. IL employs a heuristic exhaustive search to retrospectively identify the most effective model partition and exit point.

关键词： Deep learning Adaptation models Computational modeling Computer architecture Mathematical models partitioning algorithms Servers Long short term memory Autonomous vehicles Load modeling

来源：评论

学校读者我要写书评

暂无评论

A Heuristic Algorithm for Recompressing Compressed Data Files

A Heuristic Algorithm for Recompressing Compressed Data File...

引用

Data Compression Conference (DCC)

作者： Bekir Tevfik Akgün Yeditepe University Turkey

ISBN: (数字)9798331534714

ISBN: (纸本)9798331534721

We proposed a recompression methodology based on partitioning of a compressed data file into fixed-sized pages. We select some pages to form a separate file called the sub-file showing in the Figure. After saving the encoded addressing information related to these selected pages and the remaining pages of the input file in an output file, we then compress the newly created sub-file by using an off-the-shelf compression program as 7-Zip in ultra-option. The process ends by adding the compressed sub-file content to the output file. The recompression will be successfully done if the reduction amount obtained from the sub-file compression is higher than the size of the encoded addressing information. We partition an input file into $m$ pages, therefore, there are $2^{m}$ different sub-files to be formed, compressed and tested. In this work, we propose a heuristic algorithm, which is based on the selection of such a page containing identical data words, for determining pages to form a sub-file. We need only three experiments instead of $2^{m}$ experiments due to three data word versions of the algorithm. The time complexity of the algorithm including a linear searching is $O(n)$ , where $n$ is the size of the input file in bytes. We also proposed an encoding schema to store the addressing information related to the selected pages with minimal memory cost, by designing a hierarchical partition tree of the input file in three levels [1]. Our results showed that we are able to obtain Space Saving rates more than 0.01% on chosen example compressed data files showing a few in the Table, while obtaining 3.3% for one case.

关键词： Costs Heuristic algorithms Data compression Encoding partitioning algorithms Time complexity

来源：评论

学校读者我要写书评

暂无评论

Dataflow-Aware Macro Placement Based on Simulated Evolution Algorithm for Mixed-Size Designs

引用

IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS 2021年第5期29卷 973-984页

作者： Lin, Jai-Ming Deng, You-Lun Yang, Ya-Chu Chen, Jia-Jian Lu, Po-Chen Natl Cheng Kung Univ Dept Elect Engn Tainan 70101 Taiwan

This article proposes a novel approach to handle macro placement. Previous works usually apply the simulated annealing (SA) algorithm to handle this problem. However, the SA-based approaches usually have difficulty in handling preplaced macros and require longer runtime. To resolve these problems, we propose a macro placement procedure based on the corner stitching data structure and then apply an efficient and effective simulated evolution algorithm to further refine placement results. In order to relieve local routing congestion, we propose to expand areas of movable macros according to the design hierarchy before applying the macro placement algorithm. Finally, we extend our macro placement methodology to consider dataflow constraint so that dataflow-related macros can be placed at close locations. The experimental results show that our approach obtains a better solution than a previous macro placement algorithm and a tool. Besides, placement quality can be further improved when the dataflow constraint is considered.

关键词： Standards Routing partitioning algorithms DH-HEMTs Data structures Very large scale integration Tools Design hierarchy (DH) macro placement mixed size physical design simulated evolution

来源：评论

学校读者我要写书评

暂无评论

A Fuzzy C-Means Clustering Algorithm for Real Medical Image Segmentation

A Fuzzy C-Means Clustering Algorithm for Real Medical Image ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Feifei Zhang Fei Shi Dayong Ren Yue Li School of Computer Science and Technology Xinjiang University Urumqi China Key Laboratory of Signal Detection and Processing Xinjiang University Urumqi China National Key Laboratory for Novel Software Technology Nanjing University Nanjing China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In the field of medical image processing, the presence of noise can often result in the obfuscation of crucial details, which in turn can have a detrimental impact on the accuracy of clinical diagnoses. In order to effectively remove noise and improve segmentation performance, this paper proposes a refined Fuzzy C-Means (FCM) algorithm, designated as MKL-FCM. The method commences with Poisson denoising, which enables more effective handling of the noise characteristics inherent to medical images. Subsequently, multi-scale Kullback-Leibler divergence is utilised to analyse local image information across varying scales, facilitating the differentiation between tissues and pathological regions. Furthermore, tight wavelet frames are capable of capturing fine image details, while reverse optimisation of the objective function serves to correct errors in feature reconstruction, thereby enhancing the accuracy of the segmentation process. The experimental results demonstrate that MKL-FCM enhances both image clarity and segmentation accuracy, and outperforms existing methods in terms of efficiency.

关键词： Image segmentation Accuracy Noise reduction Noise Signal processing algorithms Clustering algorithms Linear programming partitioning algorithms Speech processing Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

An Efficient Joint Angle and Range Estimation Scheme for THz ISAC Systems with Single RF Chain

An Efficient Joint Angle and Range Estimation Scheme for THz...

引用

IEEE Conference on Wireless Communications and Networking

作者： Chenhao Luo Aimin Tang UM-SJTU Joint Institute Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

Integrated sensing and communications (ISAC) has attracted great attention in recent years. Thanks to the large available bandwidth and large antenna array, terahertz (THz) ISAC has the potential to achieve high range and angle resolution for wireless sensing. Considering the cost and complexity of implementing THz radio frequency (RF) chains, analog beamforming with a single RF chain is a more practical choice. However, due to the extremely narrow THz beamwidth, the time-division sweeping for angle estimation may lead to unacceptable overhead. Therefore, how to achieve efficient angle estimation under a single RF chain is a challenging problem. In this paper, we leverage the ThzPrism architecture to enhance the angle coverage, which can spread the beams at different subcarriers in different directions. We show that the angle and range estimations are coupled under the ThzPrism architecture. To this end, a joint angle and range estimation problem is formulated. By utilizing the sparse feature of target distribution, a multi-measurement vector (MMV) compressed sensing algorithm is proposed to resolve this problem. Moreover, to reduce the complexity of the algorithm, we leverage the beam-spreading feature of ThzPrism to partition the whole angle coverage area into many small subareas, which can effectively reduce the dimension of the original compressed sensing problem. The effectiveness of our proposed scheme is validated by simulation results, which show a better performance of both angle and range estimations than the existing scheme.

关键词： Radio frequency Wireless sensor networks Simulation Estimation Integrated sensing and communication partitioning algorithms Complexity theory Compressed sensing Terahertz communications Antenna arrays

来源：评论

学校读者我要写书评

暂无评论

CritiPrefill: A Segment-wise Criticality-based Approach for Prefilling Acceleration in LLMs

CritiPrefill: A Segment-wise Criticality-based Approach for ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Junlin Lv Yuan Feng Xike Xie Xin Jia Qirong Peng Guiming Xie AI Center Guangdong OPPO Mobile Telecommunications Corp.Ltd China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Large language models have achieved notable success across various domains, yet efficient inference is still limited by the quadratic computation complexity of the attention mechanism. The inference consists of prefilling and decoding phases. Although several attempts have been made to accelerate decoding, the inefficiency of the prefilling phase, especially for long-context tasks, remains a challenge. In this paper, we observe a locality in query criticality during the prefilling phase of long-context processing: adjacent query tokens tend to focus on similar subsets of the past Key-Value (KV) cache. Based on this observation, we propose CritiPrefill, a criticality-based segment-wise prefilling method. This method partitions the input sequence’s queries and KV cache into segments and blocks, utilizing a segment-wise algorithm to estimate the query criticality. By pruning noncritical computations between query segments and cache blocks in the self-attention mechanism, the prefilling process can be significantly accelerated. Extensive evaluations on multiple long-context datasets show up to 2.7x speedup on Llama3-8B and 3.0x speedup on Yi-9B for 128K context length on a single A100 GPU, with minimal quality degradation.

关键词： Large language models Signal processing algorithms Graphics processing units Estimation Signal processing Decoding Computational efficiency partitioning algorithms Speech processing Load modeling

来源：评论

学校读者我要写书评

暂无评论

Research on Efficient Processing Algorithm for Engineering Cost Data Using Cloud Computing Platform

Research on Efficient Processing Algorithm for Engineering C...

引用

Cybersecurity, Internet of Things and Soft Computing (CITSC), Asia-Europe Conference on

作者： Dantong Li Dalian University of Finance and Economics Dalian China

ISBN: (数字)9798331504205

ISBN: (纸本)9798331504212

This paper proposes a research scheme for efficient processing algorithm of engineering cost data based on cloud computing platform, aiming to improve data processing efficiency by utilizing the high-performance computing capability and distributed characteristics of cloud computing. The algorithm design takes distributed compression algorithm as the core, combines data partitioning, parallel computing and storage optimization technology to achieve efficient storage and fast access to large-scale cost data. In the system implementation, a parallel algorithm architecture that adapts to multi-node environment is designed. The algorithm performance is evaluated by building a simulation model. The experimental results show that the method proposed in this paper is superior to traditional algorithms in data compression rate, processing speed and resource utilization. When processing 1TB cost data, the compression rate reaches 85%, and the processing efficiency is improved by about 60%. Data analysis further verifies the stability and scalability of the algorithm, and provides theoretical basis and technical support for intelligent data processing in the field of engineering cost. This study opens up a new path for engineering cost data processing, and the proposed algorithm and platform architecture can be widely used in other complex data scenarios.

关键词： Cloud computing Costs Scalability Distributed databases Computer architecture Stability analysis partitioning algorithms Resource management Compression algorithms System implementation

来源：评论

学校读者我要写书评

暂无评论

Enhancing Community Detection with Weighted Louvain

Enhancing Community Detection with Weighted Louvain

引用

International Conference on Web Research (ICWR)

作者： Akram Karimi Zarandi Ali Kamandi School of Engineering Science College of Engineering University of Tehran Iran

ISBN: (数字)9798331508913

ISBN: (纸本)9798331508920

Community detection in complex networks plays a crucial role in analyzing data structures. The Louvain algorithm is a widely used method for community detection; however, it can be improved by incorporating edge weight information. In this study, we introduce a weighted version of the Louvain algorithm and compare it with the standard Louvain, Leiden, and Girvan-Newman methods. Experiments were conducted on six networks, including three small graphs (fewer than 500 nodes) and three large graphs (approximately 2500 nodes). The results show that the weighted Louvain algorithm not only achieves higher modularity than the original Louvain and other methods, but also reduces the execution time compared to the original Louvain. This improvement is attributed to leveraging edge weight information, which leads to faster convergence and more precise network partitioning.

关键词： Accuracy Image edge detection Heuristic algorithms Scalability Clustering algorithms Complex networks Approximation algorithms partitioning algorithms Standards Optimization

来源：评论

学校读者我要写书评

暂无评论

Representative Community Detection algorithms for Attribute-Free Networks

Representative Community Detection Algorithms for Attribute-...

引用

International Conference on Computer and Automation Engineering, ICCAE

作者： Dongming Chen Xin Zou Yuxing He Mingzhao Xie Dongqi Wang Software College Northeastern University Shenyang China

ISBN: (数字)9798331533816

ISBN: (纸本)9798331533823

In the field of network analysis, methods for identifying community structure often involve optimizing a specific objective function to achieve a single optimal allocation from network nodes to communities. However, in practice, we often encounter multiple division schemes with high quality scores that are close to the overall optimum. In fact, an accurate depiction of the community structure is more appropriately achieved by a series of high-quality division schemes rather than relying on a single optimal solution alone. However, such a collection of network divisions may be difficult to interpret, as its size may rapidly expand to hundreds or even thousands. To this end, this paper introduces an innovative strategy to reveal the diversity of network community structures more comprehensively and to provide a set of more globally-visioned delineation results by clustering similar network divisions and then selecting representative divisions from each cluster.

关键词： Clustering methods Network analyzers Complex networks Network coding Linear programming partitioning algorithms Resource management Object recognition Detection algorithms Periodic structures

来源：评论

学校读者我要写书评

暂无评论

Sum-Rate Maximization for Partitioned-IRS-Assisted Hybrid NOMA Networks

Sum-Rate Maximization for Partitioned-IRS-Assisted Hybrid NO...

引用

IEEE Conference on Wireless Communications and Networking

作者： Songchao Chen Fang Liu Yuanan Liu School of Electronic Engineering Beijing University of Posts and Telecommunications Beijing China

ISBN: (数字)9798350368369

ISBN: (纸本)9798350368376

Intelligent Reflecting Surfaces (IRS) have emerged as a revolutionary solution in next-generation (6G) wireless communication systems. Comprising numerous programmable reflecting units, IRS technology improves data transmission rates by dynamically adjusting the phase shifts of reflected signals. Non-orthogonal multiple access (NOMA) is a technique that can greatly enhance spectral efficiency and system sum-rate. In this paper, we consider a partitioned-IRS-assisted NOMA system in which a base station (BS) transmits superposed downlink signals to multiple users. Firstly, we establish the system sum-rate maximization problem of optimal joint IRS phase shifts and the partition sub-surfaces matching scheme by adapting combined channel strength(CSS)-based hierarchical pairing algorithm. However, since the formulated problem is non-convex, we use an alternating optimization (AO) technique to decompose the non-convex optimization problem into two subproblems for iterative solving. To further optimize the sum-rate, we also propose a NOMA user selection criteria based on the minimum signal-to-noise ratio (SNR) difference. The simulation results show a significant improvement in the performance of the IRS-assisted NOMA system compared to various traditional benchmarking schemes.

关键词： Wireless communication NOMA Spectral efficiency System performance Simulation Downlink partitioning algorithms Iterative methods Optimization Signal to noise ratio

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：