检索结果-内蒙古大学图书馆

Lightweight Task-Oriented Semantic Communication Empowered by Large-Scale AI Models

IEEE Transactions on Vehicular technology 2025年

作者： Liu, Chuanhong Guo, Caili Yang, Yang Chen, Mingzhe Quek, Tony Q. S. Beijing University of Posts and Telecommunications Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing100876 China Beijing University of Posts and Telecommunications Beijing Laboratory of Advanced Information Networks School of Information and Communication Engineering Beijing100876 China University of Miami Department of Electrical and Computer Engineering Institute for Data Science and Computing Coral GablesFL United States Singapore University of Technology and Design Dept. of Information Systems Technology and Design 487372 Singapore

Recent studies have focused on leveraging large-scale artificial intelligence (LAI) models to improve semantic representation and compression capabilities. However, the substantial computational demands of LAI models pose significant challenges for real-time communication scenarios. To address this, this paper proposes utilizing knowledge distillation (KD) techniques to extract and condense knowledge from LAI models, effectively reducing model complexity and computation latency. Nevertheless, the inherent complexity of LAI models leads to prolonged inference times during distillation, while their lack of channel awareness compromises the distillation performance. These limitations make standard KD methods unsuitable for task-oriented semantic communication scenarios. To address these issues, we propose a fast distillation method featuring a pre-stored compression mechanism that eliminates the need for repetitive inference, significantly improving efficiency. Furthermore, a channel adaptive module is incorporated to dynamically adjust the transmitted semantic information based on varying channel conditions, enhancing communication reliability and adaptability. In addition, an information bottleneck-based loss function is derived to guide the fast distillation process. Simulation results verify that the proposed scheme outperform baselines in term of task accuracy, model size, computation latency, and training data requirements. © 1967-2012 IEEE.

关键词： NP-hard

来源：评论

学校读者我要写书评

暂无评论

Leveraging LLMs for Utility-Focused Annotation: Reducing Manual Effort for Retrieval and RAG

arXiv

引用

arXiv 2025年

作者： Zhang, Hengran Tang, Minghao Bi, Keping Guo, Jiafeng Liu, Shihao Shi, Daiting Yin, Dawei Cheng, Xueqi CAS Key Lab of Network Data Science and Technology ICT CAS University of Chinese Academy of Sciences Beijing China Baidu Inc Beijing China Nankai University Tianjin China

Retrieval models typically rely on costly human-labeled query-document relevance annotations for training and evaluation. To reduce this cost and leverage the potential of Large Language Models (LLMs) in relevance judgments, we aim to explore whether LLM-generated annotations can effectively replace human annotations in training retrieval models. Retrieval usually emphasizes relevance, which indicates "topic-relatedness" of a document to a query, while in RAG, the value of a document (or utility), depends on how it contributes to answer generation. Recognizing this mismatch, some researchers use LLM performance on downstream tasks with documents as labels, but this approach requires manual answers for specific tasks, leading to high costs and limited generalization. In another line of work, prompting LLMs to select useful documents as RAG references eliminates the need for human annotation and is not task-specific. If we leverage LLMs’ utility judgments to annotate retrieval data, we may retain cross-task generalization without human annotation in large-scale corpora. Therefore, we investigate utility-focused annotation via LLMs for large-scale retriever training data across both in-domain and out-of-domain settings on the retrieval and RAG tasks. To reduce the impact of low-quality positives labeled by LLMs, we design a novel loss function, i.e., Disj-InfoNCE. Our experiments reveal that: (1) Retrievers trained on utility-focused annotations significantly outperform those trained on human annotations in the out-of-domain setting on both tasks, demonstrating superior generalization capabilities. (2) LLM annotation does not replace human annotation in the in-domain setting. However, incorporating just 20% human-annotated data enables retrievers trained with utility-focused annotations to match the performance of models trained entirely with human annotations, while adding 100% human annotations further significantly enhances performance on both tasks. We hope our wor

关键词： Human engineering

来源：评论

学校读者我要写书评

暂无评论

DAMBLO: Improving arrhythmia classification with plug-and-play dual attention-based multiscale feature learning blocke

引用

Expert Systems with Applications 2025年 285卷

作者： Zhuang, Tianming Qin, Zhiguang You, Li Deng, Erqiang Ding, Yi Cao, Mingsheng Guo, Yingkun Network and Data Security Key Laboratory of Sichuan Province University of Electronic Science and Technology of China Sichuan Chengdu610054 China School of Information and Software Engineering University of Electronic Science and Technology of China Sichuan Chengdu610054 China Department of Molecular Genetics Erasmus University Medical Center Rotterdam Netherlands Department of Radiology Key Laboratory of Obstetric and Gynecologic and Pediatric Diseases and Birth Defects of Ministry of Education West China Second University Hospital Sichuan University Sichuan Chengdu610072 China

The electrocardiogram diagnosis played an important role in early arrhythmia prevention and cardiovascular disease detection. How to analysis and detect the electrocardiogram automatically becomes a challenging task in clinical practice. Convolutional neural network based approaches have been widely applied in the field of automatic arrhythmia detection and analysis. However, the existing methods fails to preserve the abundant and essential latent representatives from multiple dimensions. To address these problems, a dual spatial-channel aggregation module is proposed, which expects to concentrate on the prominent target information along spatial and channel dimension of electrocardiogram signals. Moreover, a multi-scale enhancement module with channel-wise gating mechanism is also proposed to enlarge the receptive resolution and further to improve the learning ability of long-range dependencies from the global information. Furthermore, these two proposed modules are integrated as a "plug-and-play" block that can be easily adopted as a drop-in replacement for any 3×3 convolution layers in convolutional neural networ based architectures. In order to evaluate the effectiveness and generalization of the proposed block, it is plugged in a series of convolutional neural network based frameworks and are evaluated on two benchmark datasets, MIT-BIH and PTB. The extensive experimental results show that the proposed "plug-and-play" block can significantly improve the arrhythmia classification performance. Moreover, compared to other state-of-the-art methods, the framework plugged with the proposed block can achieve a competitive classification performance, proving the feasibility and advancement of the proposed dual attention-based multiscale feature learning block. © 2025 Elsevier Ltd

关键词： Electrocardiography

来源：评论

学校读者我要写书评

暂无评论

Magnetic fluid enabled hexagonal fiber grating for vector magnetic field sensing

引用

Photonics Research 2025年第6期13卷 1726-1735页

作者： Siyu Chen Chen Jiang Yuehui Ma Yunhe Zhao Lilong Dai Qianqian Huang Wei Chen Chengbo Mou Yunqi Liu Key Laboratory of Specialty Fiber Optics and Optical Access Networks Joint International Research Laboratory of Specialty Fiber Optics and Advanced Communication Shanghai Institute for Advanced Communication and Data Science Shanghai University Shanghai 200444 China Advanced Photonic Technology Lab College of Electronics and Optical Engineering Nanjing University of Posts and Telecommunications Nanjing 210023 China Institute of Logistics Science and Engineering Shanghai Maritime University Shanghai 201306 China e-mail: mouc1@***

Optical fiber magnetic field sensors play a crucial role in aerospace and medical fields due to their high sensitivity, fast response time, and resistance to electromagnetic interference. Most current research primarily focuses on detecting magnetic field intensity; however, the magnetic field is a vector field with both intensity and direction, making vector magnetic field measurement significantly important in various fields. Here, we experimentally demonstrated a vector magnetic field sensor using a magnetic fluid (MF) enabled hexagonal fiber grating. Such a specialty optical fiber device features strong asymmetric evanescent field distribution along the index perturbed area, from which the overcoated MF can sense the external magnetic field. When the fiber magnetometer operated at the dispersion turning point, a maximum sensitivity of 10.48 nm/mT was achieved within a range of 0–20.7 mT, which is one order of magnitude greater than that of conventional fiber grating sensors. Utilizing the polygon optical fiber, our demonstrated device simultaneously achieves a maximum orientation sensitivity of 1.17 nm/deg within a range of 0°–30°. This hexagonal fiber grating as an excellent vector magnetic field sensor may be used in military, aerospace, medical sectors, etc.

关键词： Effective refractive index Fiber Bragg gratings Fiber optic sensors Long period fiber gratings Medical fields Optical fibers

来源：评论

学校读者我要写书评

暂无评论

2D-DOA Estimation for Polarized Massive Antenna Array: Geometric and Algorithmic Guidelines

引用

IEEE Transactions on Vehicular technology 2025年

作者： Wen, Fangqing Wang, Han Gui, Guan Yuen, Chau Sari, Hikmet Adachi, Fumiyuki China Three Gorges University Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering Yichang443002 China Hubei University of Automotive Technology Institute of Vehicle Information Control and Network Technology Shiyan442002 China Yichun University College of Physical Sciences and Engineering Technology Yichun336000 China City University of Macau Faculty of Data Science China Nanjing University of Posts and Telecommunications College of Telecommunications and Information Engineering Nanjing210003 China Nanyang Technological University School of Electrical and Electronics Engineering 639798 Singapore Sendai980-8572 Japan

The utilization of an Electromagnetic Vector Sensor (EMVS) array presents valuable insights in wireless communications and radars by incorporating the advantages of spatial, temporal, and polarization diversities. This study focuses on addressing the problem of estimating the Two-Dimensional (2D) Direction-of-Arrival (DOA) for an arbitrarily spaced EMVS array. Firstly, the Cramer-Rao Bounds (CRB) for estimating the 2D-DOA and polarized parameters are derived, which provides design guideline for array geometry. As suggested by the CRB, a sparse planar EMVS array is a more appealing choice compared to the popular coprime EMVS array. Thereafter, a Parallel Factor (PARAFAC)-based weighted phase fitting algorithm is presented, designed to fully explore the aperture of the conformal array as well as capitalize on the tensorial structure inherent in array data. The proposed PARAFAC algorithm not only provides a closed-form solution for conformal arrays but also exhibits superior accuracy in 2D-DOA estimation compared to existing approaches. Numerical results corroborate the theoretical advantages of the proposed array geometry and PARAFAC algorithm. © 1967-2012 IEEE.

关键词： Cramer-Rao bounds

来源：评论

学校读者我要写书评

暂无评论

Critical (P5, W4)-Free Graphs

arXiv

引用

arXiv 2025年

作者： Xia, Wen Jooken, Jorik Goedgebeur, Jan Beaton, Iain Cameron, Ben Huang, Shenwei College of Computer Science Nankai University Tianjin300071 China Tianjin Key Laboratory of Network and Data Security Technology Nankai University Tianjin300071 China Department of Computer Science KU Leuven Campus Kulak-Kortrijk Kortrijk8500 Belgium Department of Applied Mathematics Computer Science and Statistics Ghent University Ghent9000 Belgium Department of Mathematics and Statistics Acadia University WolfvilleNS Canada School of Mathematical and Computational Sciences University of Prince Edward Island CharlottetownPE Canada

A graph G is k-vertex-critical if χ(G) = k but χ(G − v) 1, H2)-free if it contains no induced subgraph isomorphic to H1 nor H2. A W4 is the graph consisting of a C4 plus an additional vertex adjacent to all the vertices of the C4. We show that there are finitely many k-vertex-critical (P5, W4)-free graphs for all k ≥ 1 and we characterize all 5-vertex-critical (P5, W4)-free graphs. Our results imply the existence of a polynomial-time certifying algorithm to decide the k-colorability of (P5, W4)-free graphs for each k ≥ 1 where the certificate is either a k-coloring or a (k + 1)-vertex-critical induced subgraph. Copyright © 2025, The Authors. All rights reserved.

关键词： Graph algorithms

来源：评论

学校读者我要写书评

暂无评论

Extending attributes page:a scheme for enhancing the reliability of storage system metadata

引用

Journal of Zhejiang University-science A(Applied Physics & Engineering) 2009年第8期10卷 1106-1113页

作者： Juan WANG Dan FENG Fang WANG Cheng-tao LU MOE Key Lab of Data Storage System Huazhong University of Science and Technology Wuhan 430074 China

In an object-based storage system,a novel scheme named EAP(extending attributes page) is presented to enhance the metadata reliability of the system by adding the user object file information attributes page for each user object and storing the file-related attributes of each user object in object-based storage *** EAP scheme requires no additional hardware equipments compared to a general method which uses backup metadata servers to improve the metadata *** a Markov chain,this paper compares the metadata reliability of the system using the EAP scheme with that using only metadata servers to offer the file metadata *** results demonstrate that the EAP scheme can dramatically enhance the reliability of storage system metadata.

关键词： Metadata server Metadata reliability Markov model Extending attributes page （EAP）

来源：评论

学校读者我要写书评

暂无评论

MIA-UNet:Multi-Scale Iterative Aggregation U-network for Retinal Vessel Segmentation

引用

Computer Modeling in Engineering & sciences 2021年第11期129卷 805-828页

作者： Linfang Yu Zhen Qin Yi Ding Zhiguang Qin Network and Data Security Key Laboratory of Sichuan Province University of Electronic Science and Technology of ChinaChengdu610054China

As an important part of the new generation of information technology,the Internet of Things(IoT)has been widely concerned and regarded as an enabling technology of the next generation of health care *** fundus photography equipment is connected to the cloud platform through the IoT,so as to realize the realtime uploading of fundus images and the rapid issuance of diagnostic suggestions by artificial *** the same time,important security and privacy issues have *** data uploaded to the cloud platform involves more personal attributes,health status and medical application data of *** leaked,abused or improperly disclosed,personal information security will be ***,it is important to address the security and privacy issues of massive medical and healthcare equipment connecting to the infrastructure of IoT healthcare and health *** meet this challenge,we propose MIA-UNet,a multi-scale iterative aggregation U-network,which aims to achieve accurate and efficient retinal vessel segmentation for ophthalmic auxiliary diagnosis while ensuring that the network has low computational complexity to adapt to mobile *** this way,users do not need to upload the data to the cloud platform,and can analyze and process the fundus images on their own mobile terminals,thus eliminating the leakage of personal ***,the interconnection between encoder and decoder,as well as the internal connection between decoder subnetworks in classic U-Net are redefined and ***,we propose a hybrid loss function to smooth the gradient and deal with the imbalance between foreground and *** with the UNet,the segmentation performance of the proposed network is significantly improved on the premise that the number of parameters is only increased by 2%.When applied to three publicly available datasets:DRIVE,STARE and CHASE DB1,the proposed network achieves the accuracy/F1-score of 96.33%/84.34%

关键词： Retinal vessel segmentation security and privacy redesigned skip connection feature maps aggregation hybrid loss function

来源：评论

学校读者我要写书评

暂无评论

Exploring the Impact of Personality Traits on LLM Bias and Toxicity

arXiv

引用

arXiv 2025年

作者： Wang, Shuo Li, Renhao Chen, Xi Wong, Derek F. Yuan, Yulin Yang, Min NLP2CT Lab Department of Computer and Information Science University of Macau China Shenzhen Key Laboratory for High Performance Data Mining Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China Linguistics and Multilingual Studies Nanyang Technological University Singapore Department of Chinese Language and Literature University of Macau China

With the different roles that AI is expected to play in human life, imbuing large language models (LLMs) with different personalities has attracted increasing research interests. While the "personification" enhances human experiences of interactivity and adaptability of LLMs, it gives rise to critical concerns about content safety, particularly regarding bias, sentiment and toxicity of LLM generation. This study explores how assigning different personality traits to LLMs affects the toxicity and biases of their outputs. Leveraging the widely accepted HEXACO personality framework developed in social psychology, we design experimentally sound prompts to test three LLMs’ performance on three toxic and bias benchmarks. The findings demonstrate the sensitivity of all three models to HEXACO personality traits and, more importantly, a consistent variation in the biases, negative sentiment and toxicity of their output. In particular, adjusting the levels of several personality traits can effectively reduce bias and toxicity in model performance, similar to humans’ correlations between personality traits and toxic behaviors. The findings highlight the additional need to examine content safety besides the efficiency of training or fine-tuning methods for LLM personification. They also suggest a potential for the adjustment of personalities to be a simple and low-cost method to conduct controlled text generation. © 2025, CC BY.

关键词： Social psychology

来源：评论

学校读者我要写书评

暂无评论

Trusted Clustering Based Federated Learning in Edge networks

引用

IEEE Transactions on Mobile Computing 2025年

作者： Liu, Yi-Jing Zhang, Long Li, Xiaoqian Du, Hongyang Feng, Gang Qin, Shuang Wang, Jiacheng University of Electronic Science and Technology of China National Key Lab on Wireless Communications Chengdu China Tsinghua University Sichuan Energy Internet Research Institute Chengdu China University of Hong Kong Department of Electrical and Electronic Engineering Hong Kong Hong Kong Nanyang Technological University College of Computing and Data Science Singapore

Federated learning (FL) is integral to advancing edge intelligence by enabling collaborative machine learning. In FL-empowered edge networks, computing nodes first train local models and then send them to an or multiple aggregation node(s) for global model collaboration. However, the trustworthiness of both local and global models in conventional FL frameworks is compromised due to inadequate model security and transparency. Distributed ledger technique (DLT) can address this issue by leveraging multi-nodes trust capabilities to support distributed consensus. However, model training and consensus performance of DLT may significantly degrade due to instability and resource constraints of edge networks. Sharding technique provides an effective approach by dividing the ledger into smaller and manageable shards. In this paper, to improve model training and consensus performance, we propose a trusted FL framework by incorporating sharding DLT into FL frameworks. We construct a theoretical model to investigate the relationship between model training performance, consensus efficiency, and capacity of edge nodes regarding storage, computing and communications. Based on the theoretical model, we propose a trusted clustering scheme to aggregate local models. Numerical results show that our proposed scheme significantly improves network throughput for transmitting models while guaranteeing model learning performance in comparison with some classical baselines. © 2002-2012 IEEE.

关键词： Distributed ledger

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：