检索结果-内蒙古大学图书馆

23rd USENIX Conference on File and Storage Technologies, FAST 2025

作者： Chen, Weijian He, Shuibing Qu, Haoyang Zhang, Ruidong Yang, Siling Chen, Ping Zheng, Yi Huai, Baoxing Chen, Gang The State Key Laboratory of Blockchain and Data Security Zhejiang University China Institute of Blockchain and Data Security China Zhejiang Key Laboratory of Big Data Intelligent Computing China Huawei Cloud China

ISBN: (纸本)9781939133458

Modern advanced large language model (LLM) applications often prepend long contexts before user queries to improve model output quality. These contexts frequently repeat, either partially or fully, across multiple queries. Existing systems typically store and reuse the keys and values of these contexts (referred to as prefix KVs) to reduce redundant computation and time to first token (TTFT). When prefix KVs need to be stored on disks due to insufficient CPU memory, reusing them does not always reduce TTFT, as disk I/O latency is high. In this paper, we propose IMPRESS, an importance-informed multi-tier prefix KV storage system to reduce I/O delay for LLM inference by only loading important prefix KVs. IMPRESS first leverages the insight that there is significant similarity in important token index sets across attention heads and introduces an I/O-efficient important KV identification algorithm. It then optimizes prefix KV storage and caching through importance-informed KV management, reducing TTFT during model inference. Our experimental results show that IMPRESS can reduce TTFT by up to 2.8× compared to state-of-the-art systems, while maintaining comparable inference accuracy. © 2025 FAST. All Rights Reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Underwater target recognition based on adaptive multi-feature fusion network

引用

Multimedia Tools and Applications 2025年第10期84卷 7297-7317页

作者： Pan, Xiaoying Sun, Jia Feng, TianHao Lei, MingZhu Wang, Hao Zhang, WuXia Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing Xi’an710121 China Xi’an Key Laboratory of Big Data and Intelligent Computing Xi’an710121 China School of Computer Science and Technology Xi’an University of Post & Telecommunications Xi’an710121 China

Due to the complexity of the underwater environment, underwater acoustic target recognition is more challenging than ordinary target recognition, and has become a hot topic in the field of underwater acoustics research. In recent years, deep learning has been widely used in underwater recognition due to its powerful feature learning capabilities. However, most deep learning-based underwater target recognition methods use only the time-frequency domain features of signals, ignoring the frequency domain image features of signals. Additionally, they only use simple summation or splicing strategies to fuse and discriminate the extracted features, which leads to limited improvement in the accuracy of underwater target recognition. To address these issues, this paper proposes an underwater target recognition method based on an adaptive multi-feature fusion network. The method consists of a data pre-processing module, a multi-dimensional feature extraction module, and an adaptive multi-feature fusion module. In the multi-dimensional feature extraction module, the long short time memory network (LSTM) and the one-dimensional convolutional neural network (1DCNN) are used to extract the time-frequency features of the underwater signal. Furthermore, the two-dimensional convolutional neural network (2DCNN) is used to extract the image features in the frequency domain. The adaptive multi-feature fusion module uses an attention mechanism for adaptive weighted fusion to make full use of the learned features. The effectiveness of the proposed method is validated on the ShipEar dataset, and the recognition accuracy reaches 94.92%, which is higher than other existing methods. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A data representation method using distance correlation

引用

Frontiers of Computer Science 2025年第1期19卷 1-14页

作者： Xinyan LIANG Yuhua QIAN Qian GUO keyin ZHENG Institute of Big Data Science and Industry Shanxi UniversityTaiyuan 030006China Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education Shanxi UniversityTaiyuan 030006China School of Computer Science and Technology Taiyuan University of Science and TechnologyTaiyuan 030024China Shanxi Key Laboratory of Big Data Analysis and Parallel Computing Taiyuan University of Science and TechnologyTaiyuan 030024China

Association in-between features has been demonstrated to improve the representation ability of data. However, the original association data reconstruction method may face two issues: the dimension of reconstructed data is undoubtedly higher than that of original data, and adopted association measure method does not well balance effectiveness and efficiency. To address above two issues, this paper proposes a novel association-based representation improvement method, named as AssoRep. AssoRep first obtains the association between features via distance correlation method that has some advantages than Pearson’s correlation coefficient. Then an improved matrix is formed via stacking the association value of any two features. Next, an improved feature representation is obtained by aggregating the original feature with the enhancement matrix. Finally, the improved feature representation is mapped to a low-dimensional space via principal component analysis. The effectiveness of AssoRep is validated on 120 datasets and the fruits further prefect our previous work on the association data reconstruction.

关键词： association representation distance correlation classification

来源：评论

学校读者我要写书评

暂无评论

Granular Sequential Three-Way Decision for Specific Decision Classes

引用

IEEE Transactions on Fuzzy Systems 2025年第5期33卷 1650-1663页

作者： Cheng, Yunlong Yang, Xiuhua Zhang, Qinghua Shao, Yabin Wang, Guoyin Chongqing University of Posts and Telecommunications School of Science Chongqing400065 China Chongqing University of Posts and Telecommunications Chongqing Key Laboratory of Tourism Multisource Data Perception and Decision Ministry of Culture and Tourism Key Laboratory of Big Data Intelligence Computing Chongqing Key Laboratory of Computational Intelligence Chongqing400065 China

Sequential three-way decision (S3WD) is an efficient granular computing paradigm for dealing with uncertain problems. However, it is primarily oriented to all decision classes, which contradicts the fact that decisions are typically for the specific decision classes. Meanwhile, most S3WD models hide the topological structure of the granules, leading to difficulties in semantic interpretation. To address the issues, integrating model construction, attribute reduction and knowledge extraction, a general framework of granular sequential three-way decision (GrS3WD) for the specific decision classes is proposed to improve semantic interpretation and computational efficiency. First, a two-stage trisecting strategy and a GrS3WD model are proposed to integrate model construction with attribute reduction. Its main advantage is that it retains the topological structure of granules, which not only enhances semantic interpretation but also avoids unnecessary double counting. Second, three acceleration strategies and a novel granular sequential three-way reduction (GrS3WR) algorithm are proposed to fast obtain a classification-based reduct or a class-specific reduct. Finally, the decision rules with multi-granularity can be directly extracted from the concept tree generated by GrS3WR. Experimental results demonstrate that a class-specific reduct usually has fewer attributes and better classification performance than a classification-based reduct. Moreover, GrS3WR can significantly improve the computational efficiency of attribute reduction. © 1993-2012 IEEE.

关键词： Granular computing

来源：评论

学校读者我要写书评

暂无评论

Improved quantum image weighted average filtering algorithm

引用

Quantum Information Processing 2025年第5期24卷 1-24页

作者： Yuan, Suzhen Li, Xianli Yinxia, Shu Qing, Xianrong Deng, Jermiah D. College of Opto Electronic Engineering Chongqing University of Posts and Telecommunications Chongqing Chongqing400065 China School of Computing University of Otago Dunedin Dunedin9054 New Zealand School of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing Chongqing400065 China Chongqing Key Laboratory of Computational Intelligence Key Laboratory of Big Data Intelligent Computing Key Laboratory of Cyberspace Big Data Intelligent Security Ministry of Education Chongqing University of Posts and Telecommunications Chongqing Chongqing400065 China

Average filtering plays a vital role in image smoothing tasks. However, existing quantum image weighted average filtering methods suffer from high circuit complexity. Therefore, this paper proposes an improved quantum color image weighted average filtering algorithm and its corresponding quantum circuit. First, we improve the quantum circuit to prepare classical color images into a quantum state. Then, an improved quantum divider is developed, and a weighted average filter is constructed using basic quantum image processing modules. Next, to enhance the universality of the filter, a quantum comparator with lower circuit complexity is used to design a noise detection module for distinguishing noise from real signals. Finally, a quantum circuit for color image weighted average filtering is designed, and simulations are conducted on the IBM Quantum Experience (IBM Q) platform to verify the feasibility of our algorithm. The analysis shows that compared with existing methods, this method significantly reduces the circuit complexity and has better filtering performance. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Towards Adaptive Masked Structural Learning for Graph-Level Clustering

引用

IEEE Transactions on Network Science and Engineering 2025年第3期12卷 2021-2032页

作者： Yang, Jinbin Cai, Jinyu Zhang, Yunhe Huang, Sujia Wang, Shiping Fuzhou University College of Computer and Data Science Fuzhou350116 China Bank of Quanzhou Company Ltd. Quanzhou362046 China National University of Singapore Institute of Data Science 117602 Singapore Shenzhen Research Institute of Big Data Shenzhen518172 China Chinese University of Hong Kong Guangdong Provincial Key Laboratory of Big Data Computing Shenzhen518172 China

Graph data presents a vast landscape for real-world applications. Current graph-level clustering approaches predominantly utilize graph neural networks to capture the intricate structural information for graph data. However, a significant challenge arises in effectively integrate structural and feature information under the prevalent noise in the real-world scenario. The advent of masking strategies has marked significant strides in boosting model robustness, accommodating incomplete data, and enhancing generalization capabilities. Yet, research attention on leveraging mask strategy for facilitating graph-level clustering is still limited. In this paper, we introduce a novel graph-level clustering method, towards adaptive masked structural learning for graph-level clustering. The method performs adaptive masking through reconstruction loss, and jointly adaptive mask representation learning and clustering in an end-to-end unsupervised framework. The mutual information between maximized the entire graph and substructure representations is also utilized to learn to generate cluster-oriented graph-level representations. Extensive experiments on eight real graph-level benchmark datasets demonstrate the effectiveness and superiority of the proposed method. © 2013 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Retrieval-Augmented Fine-Tuning for Improving Retrieve-and-Edit Based Assertion Generation

引用

IEEE Transactions on Software Engineering 2025年第05期51卷 1591-1614页

作者： Li, Hongyan Sun, Weifeng Yan, Meng Xu, Ling Li, Qiang Zhang, Xiaohong Zhang, Hongyu Key Laboratory of Dependable Service Computing in Cyber Physical Society Chongqing University Ministry of Education Chongqing400044 China School of Big Data and Software Engineering Chongqing University Chongqing400044 China

Unit Testing is crucial in software development and maintenance, aiming to verify that the implemented functionality is consistent with the expected functionality. A unit test is composed of two parts: a test prefix, which drives the unit under test to a particular state, and a test assertion, which determines what the expected behavior is under that state. To reduce the effort of conducting unit tests manually, Yu et al. proposed an integrated approach (integration for short), combining information retrieval with a deep learning-based approach to generate assertions for test prefixes, and obtained promising results. In our previous work, we found that the overall performance of integration is mainly due to its success in retrieving assertions. Moreover, integration is limited to specific types of edit operations and struggles to understand the semantic differences between the retrieved focal-test (focal-test includes a test prefix and a unit under test) and the input focal-test. Based on these insights, we then proposed a retrieve-and-edit approach named EditAS to learn the assertion edit patterns to improve the effectiveness of assertion generation in our prior study. Despite being promising, we find that the effectiveness of EditAS can be further improved. Our analysis shows that: ① The editing ability of EditAS still has ample room for improvement. Its performance degrades as the edit distance between the retrieval assertion and ground truth increases. Specifically, the average accuracy of EditAS is 12.38% when the edit distance is greater than 5. ② EditAS lacks a fine-grained semantic understanding of both the retrieved focal-test and the input focal-test themselves, which leads to many inaccurate token modifications. In particular, an average of 25.57% of the incorrectly generated assertions that need to be modified are not modified, and an average of 6.45% of the assertions that match the ground truth are still modified. Thanks to pre-trained models employing

关键词： Software design

来源：评论

学校读者我要写书评

暂无评论

T-Assess: An Efficient data Quality Assessment System Tailored for Trajectory data 51st

T-Assess: An Efficient Data Quality Assessment System Tailor...

引用

51st International Conference on Very Large data Bases, VLDB 2025

作者： Zhu, Junhao Wang, Tao Hu, Danlei Fang, Ziquan Chen, Lu Gao, Yunjun Li, Tianyi Jensen, Christian S. Zhejiang University China Zhejiang University Zhejiang Key Laboratory of Big Data Intelligent Computing China Aalborg University Denmark

With the widespread use of GPS-enabled devices and services, trajectory data fuels services in a variety of fields, such as transportation and smart cities. However, trajectory data often contains errors stemming from inaccurate GPS measurements, low sampling rates, and transmission interruptions, yielding low-quality trajectory data with negative effects on downstream services. Therefore, a crucial yet tedious endeavor is to assess the quality of trajectory data, serving as a guide for subsequent data cleaning and analyses. Despite some studies addressing general-purpose data quality assessment, no studies exist that are tailored specifically for trajectory data. To more effectively diagnose the quality of trajectory data, we propose T-Assess, an automated trajectory data quality assessment system. T-Assess is built on three fundamental principles: i) extensive coverage, ii) versatility, and iii) efficiency. To achieve comprehensive coverage, we propose assessment criteria spanning validity, completeness, consistency, and fairness. To provide high versatility, T-Assess supports both offline and online evaluations for full-batch trajectory datasets as well as real-time trajectory streams. In addition, we incorporate an evaluation optimization strategy to achieve assessment efficiency. Extensive experiments on fourreal-life benchmark datasets offer insight into the effectiveness of T Assess at quantifying trajectory data quality beyond the capabilities of state-of-the-art data quality systems. © 2025, VLDB Endowment, All rights reserved.

关键词： data quality

来源：评论

学校读者我要写书评

暂无评论

LEAP: A Low-cost Spark SQL Query Optimizer using Pairwise Comparison 51st

LEAP: A Low-cost Spark SQL Query Optimizer using Pairwise Co...

引用

51st International Conference on Very Large data Bases, VLDB 2025

作者： Ye, Junhao Li, Jiahui Chen, Lu Mao, Yuren Gao, Yunjun Li, Tianyi Zhejiang University China Zhejiang University Zhejiang Key Laboratory of Big Data Intelligent Computing China Aalborg University Denmark

Selecting a good execution plan can significantly improve the query efficiency of Spark SQL. Several machine learning-based techniques have been proposed to select good execution plans for DBMS, but none of them perform well on Spark SQL due to the following issues.(1) Limited compatibility with Spark SQL: these approaches rely on physical operator enumeration, while Spark SQL doesn’t support it;(2) Unreliable cost estimation: they often select execution plans with poor performance due to inaccurate cost estimation;(3) Time-consuming plan enumeration: they take much time to generate a large number of candidate execution plans in Spark *** overcome these issues, in this paper, we propose LEAP, the first learned query optimizer tailored for Spark SQL, which can be integrated seamlessly into Spark SQL and solves the compatibility issue. Also, to avoid the unreliable cost value estimation, LEAP selects execution plans with an estimation-free method, which directly performs comparisons between the plans. Furthermore, LEAP employs an efficient progressive plan enumeration algorithm with pruning techniques to find better plans with fewer enumerations. Extensive experiments on three public benchmarks show the effectiveness of LEAP. It reduces the end-to-end execution time of the native optimizer by up to 54% and other learned methods by up to 94%. © 2025, VLDB Endowment, All rights reserved.

关键词： Cost benefit analysis

来源：评论

学校读者我要写书评

暂无评论

LeapGNN: Accelerating Distributed GNN Training Leveraging Feature-Centric Model Migration 23

LeapGNN: Accelerating Distributed GNN Training Leveraging Fe...

引用

23rd USENIX Conference on File and Storage Technologies, FAST 2025

作者： Chen, Weijian He, Shuibing Qu, Haoyang Zhang, Xuechen The State Key Laboratory of Blockchain and Data Security Zhejiang University China Zhejiang Lab China Institute of Blockchain and Data Security China Zhejiang Key Laboratory of Big Data Intelligent Computing China Washington State University Vancouver United States

ISBN: (纸本)9781939133458

Distributed training of graph neural networks (GNNs) has become a crucial technique for processing large graphs. Prevalent GNN frameworks are model-centric, necessitating the transfer of massive graph vertex features to GNN models, which leads to a significant communication bottleneck. Recognizing that the model size is often significantly smaller than the feature size, we propose LeapGNN, a feature-centric framework that reverses this paradigm by bringing GNN models to vertex features. To make it truly effective, we first propose a micrograph-based training strategy that leverages a refined structure to enhance locality, combined with the model migration technique, to minimize remote feature retrieval. Then, we devise a feature pre-gathering approach that merges multiple fetch operations into a single one to eliminate redundant feature transmissions. Finally, we employ a micrograph-based merging method that adjusts the number of micrographs for each worker to minimize kernel switches and synchronization overhead. Our experimental results demonstrate that LeapGNN achieves a performance speedup of up to 4.2× compared to the state-of-the-art method, namely P3 © 2025 FAST. All Rights Reserved.

关键词： Graph neural networks

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：