检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Wang, Peng Li, Zexi Zhang, Ningyu Xu, Ziwen Yao, Yunzhi Jiang, Yong Xie, Pengjun Huang, Fei Chen, Huajun Zhejiang University China Alibaba Group China Zhejiang Key Laboratory of Big Data Intelligent Computing China

Large language models (LLMs) need knowledge updates to meet the ever-growing world facts and correct the hallucinated responses, facilitating the methods of lifelong model editing. Where the updated knowledge resides in memories is a fundamental question for model editing. In this paper, we find that editing either long-term memory (direct model parameters) or working memory (nonparametric knowledge of neural network activations/representations by retrieval) will result in an impossible triangle-reliability, generalization, and locality can not be realized together in the lifelong editing settings. For long-term memory, directly editing the parameters will cause conflicts with irrelevant pretrained knowledge or previous edits (poor reliability and locality). For working memory, retrieval-based activations can hardly make the model understand the edits and generalize (poor generalization). Therefore, we propose WISE to bridge the gap between memories. In WISE, we design a dual parametric memory scheme, which consists of the main memory for the pretrained knowledge and a side memory for the edited knowledge. We only edit the knowledge in the side memory and train a router to decide which memory to go through when given a query. For continual editing, we devise a knowledge-sharding mechanism where different sets of edits reside in distinct subspaces of parameters and are subsequently merged into a shared memory without conflicts. Extensive experiments show that WISE can outperform previous model editing methods and overcome the impossible triangle under lifelong model editing of question answering, hallucination, and out-of-distribution settings across trending LLM architectures, e.g., GPT, LLaMA, and Mistral. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

IMPRESS: An Importance-Informed Multi-Tier Prefix KV Storage System for Large Language Model Inference 23

IMPRESS: An Importance-Informed Multi-Tier Prefix KV Storage...

引用

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

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

Glare-SNet: Unsupervised Glare Suppression Balance Network 27th

Glare-SNet: Unsupervised Glare Suppression Balance Network

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Li, Pei Zuo, Chengyu Wei, Wangjuan Pan, Xiaoying Wang, Zhanhao School of Computer Science and Technology Xi’an University of Posts and Telecommunications Xi’an710121 China Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing Xi’an China Shaanxi Key Laboratory of Intelligent Media Computing and interaction Xi’an China Xi’an Key Laboratory of Big Data and Intelligent Computing Xi’an China

ISBN: (纸本)9783031781681

In light of the problems associated with glare and halo effects in low-light images, as well as the inadequacy of existing processing algorithms in handling details, a glare suppression balance network based on unsupervised learning has been designed and implemented in this study. A network that perceives brightness through light decomposition has been proposed, and an illumination learning network has been designed. The idea of integration has been employed to form an unsupervised glare suppression balance network, aiming to tackle the issues of glare and the accompanying halo effects. Through objective and subjective evaluations of real images, it has been proven that the method introduced in this paper surpasses other methods in suppressing glare and enhancing the overall quality of low-light images. Furthermore, it has been shown to exhibit good performance and effectiveness in target detection. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

CQUPT-FL: Cross-Domain Sharing and Security Awareness and Early Warning Platform of Health Science big data 1

CQUPT-FL: Cross-Domain Sharing and Security Awareness and Ea...

引用

1st IEEE International Conference on Medical Artificial Intelligence, MedAI 2023

作者： Xiao, Yunpeng Zhang, Qunqing Zhao, Wanjing Li, Xufeng Peng, Jinhua Mo, Haonan Zhu, Haipeng Tang, Fei Chongqing University of Posts and Telecommunications Key Laboratory of Big Data Intelligent Computing Chongqing China

ISBN: (纸本)9798350358780

Federated learning (FL) is a rapidly growing research area in machine learning, but it is problematic. It has been questioned whether or not existing FL libraries are practical in the area of medical privacy. To address these issues, we developed the CQUPT-FL system. The system focuses on resolving the conflict between data integrity and medical data privacy protection in cross-domain and cross-institution collaborative analysis. CQUPT-FL supports distributed computing and stand-alone simulation computing methods. To deal with the problems of heterogeneity, data domain diversity, and effective data scarcity, we adopted key technologies such as multi-party secure computing and holistic information representation and studied user identification, privacy protection, and heterogeneous user alignment to achieve sustainable Cross-domain and cross-platform data fusion of letters. The goal of introducing the CQUPT-FL system is to improve the level of data privacy protection and enhance the data privacy protection mechanism, solve the machine learning dilemma in the field of medical privacy, and provide a reliable solution for cross-domain collaborative analysis. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Multi-Granular Differential Evolution Algorithm for Multi-Agent Task Allocation 9

Multi-Granular Differential Evolution Algorithm for Multi-Ag...

引用

9th International Conference on big data and Information Analytics, bigDIA 2023

作者： Gao, Peng Liu, Qun Jiang, Haihuan Chongqing University of Posts and Telecommunications Key Laboratory of Big Data Intelligent Computing Chongqing China

ISBN: (纸本)9798350330076

The multi-agent task allocation presents a fundamental challenge in the field of multi-agent systems, especially in uncertain environments. Although extensive research has been conducted on the multi-agent task allocation in deterministic environments, discussions around the multi-agent task allocation in uncertain environments are relatively scarce. In reality, uncertain data is more common in practical decision-making processes. To address the multi-agent task allocation problem in uncertain environments, this study frames it as a noisy optimization problem and proposes a novel Multi-Granular Differential Evolution (MGDE) algorithm to solve it. MGDE combines the powerful differential evolution (DE) with the granular-ball computing which has high robustness in noise. The proposed MGDE is compared with other three state-of-the-art algorithms on 12 scenarios encompassing 6 agent and task quantity combinations and 2 uncertainty levels. Experimental results demonstrate the superior performance of MGDE. © 2023 IEEE.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Few-Shot Object Detection Based On Label Constrained data Augmentation 9

Few-Shot Object Detection Based On Label Constrained Data Au...

引用

9th International Conference on big data and Information Analytics, bigDIA 2023

作者： Shi, Yan Wang, Guoyin Liu, Qun Chongqing University of Posts and Telecommunications Key Laboratory of Big Data Intelligent Computing Chongqing China

ISBN: (纸本)9798350330076

The majority of object detection methods typically depend on a significant quantity of annotated data, while few-shot object detection (FSOD) endeavors to identify novel classes of objects using a limited number of training instances. However, the limited number of samples leads to the problem of disparate data distributions between the source and target domains, which makes the generalization ability of the detector usually weak. In this paper, we combine data augmentation with fine-tuning to design a pseudo-labeled constrained model called LCDA, aiming to obtain high-quality pseudo-labels to effectively enrich the training examples. Furthermore, we leverage the pre-trained CLIP model to enhance the quality of pseudo-labels by restricting the category information as well as the designed bounding box consistency criterion. The experimental outcomes demonstrate that our model outperforms the existing models on two public datasets across various shot scenarios. The average enhancement of our method on different shots is 1.2AP, 1.6AP, 2.3AP, 2.7AP, and 4.2AP. We also validate the performance of the model in real applications of the USV dataset which shows an improvement of 1.9AP over the baseline methods. All demonstrate the effectiveness of our model. © 2023 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Edge computing Task Scheduling with Joint Blockchain and Task Caching in Industrial Internet

引用

Computers, Materials & Continua 2023年第4期75卷 2101-2117页

作者： Yanping Chen Xuyang Bai Xiaomin Jin Zhongmin Wang Fengwei Wang Li Ling School of Computer Science and Technology Xi’an University of Posts and TelecommunicationsXi’an710121China Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing Xi’an710121China Xi’an Key Laboratory of Big Data and Intelligent Computing Xi’anShaanxi710121China ZTE Corporation Shenzhen51805China

Deploying task caching at edge servers has become an effectiveway to handle compute-intensive and latency-sensitive tasks on the industrialinternet. However, how to select the task scheduling location to reduce taskdelay and cost while ensuring the data security and reliable communicationof edge computing remains a challenge. To solve this problem, this paperestablishes a task scheduling model with joint blockchain and task cachingin the industrial internet and designs a novel blockchain-assisted cachingmechanism to enhance system security. In this paper, the task schedulingproblem, which couples the task scheduling decision, task caching decision,and blockchain reward, is formulated as the minimum weighted cost problemunder delay constraints. This is a mixed integer nonlinear problem, which isproved to be nonconvex and NP-hard. To solve the optimal solution, thispaper proposes a task scheduling strategy algorithm based on an improvedgenetic algorithm (IGA-TSPA) by improving the genetic algorithm initializationand mutation operations to reduce the size of the initial solutionspace and enhance the optimal solution convergence speed. In addition,an Improved Least Frequently Used algorithm is proposed to improve thecontent hit rate. Simulation results show that IGA-TSPA has a faster optimalsolution-solving ability and shorter running time compared with the existingedge computing scheduling algorithms. The established task scheduling modelnot only saves 62.19% of system overhead consumption in comparison withlocal computing but also has great significance in protecting data security,reducing task processing delay, and reducing system cost.

关键词： Edge computing task scheduling blockchain task caching industrial security

来源：评论

学校读者我要写书评

暂无评论

Wireless Charging Scheduling for Long-term Utility Optimization

引用

ACM Transactions on Sensor Networks 2025年第1期21卷 1-31页

作者： Xu, Jia Chen, Wenbin Dai, Haipeng Xu, Lijie Xiao, Fu Liu, Linfeng Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing University of Posts and Telecommunications Nanjing China The State Key Laboratory for Novel Software Technology Nanjing University Nanjing China

Wireless power transmission has been widely used to replenish energy for wireless sensor networks, where the energy consumption rate of sensor nodes is usually time varying and indefinite. However, few works have investigated the problem of long-term charging scheduling with random variable. This article designs an optimization model for the long-term scheduling of chargers to maximize the time-averaged charging utility while ensuring its time-averaged constraints of budget and response rate. The Lyapunov optimization technique is adopted to transform the stochastic optimization problem into a deterministic optimization problem, which remains NP-hard. Thus, an approximation algorithm following greedy approach is proposed to solve the deterministic optimization problem. We further provide the theoretical analysis of feasibility and performance guarantee of the proposed scheduling algorithm. The simulation results show that our algorithm outperforms three comparison algorithms by 6.53%, 20.04%, and 19.97% in terms of time-averaged charging utility, as well as by 11.25%, 4.42%, and 3.73% in terms of time-averaged response rate on average. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Inductive power transmission

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：