检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Xiao, Mingxuan Xiao, Yan Ji, Shunhui Cai, Hanbo Xue, Lei Zhang, Pengcheng Key Laboratory of Water Big Data Technology Ministry of Water Resources College of Computer Science and Software Engineering Hohai University Nanjing211100 China School of Cyber Science and Technology Shenzhen Campus of Sun Yat-sen University Shenzhen518107 China

Benefiting from the advancements in LLMs, NLP software has undergone rapid development. Such software is widely employed in various safety-critical tasks, such as financial sentiment analysis, toxic content moderation, and log generation. Unlike traditional software, LLM-based NLP software relies on prompts and examples as inputs. Given the complexity of LLMs and the unpredictability of real-world inputs, quantitatively assessing the robustness of such software is crucial. However, to the best of our knowledge, no automated robustness testing methods have been specifically designed to evaluate the overall inputs of LLM-based NLP software. To this end, this paper introduces the first AutOmated Robustness Testing frAmework, AORTA, which reconceptualizes the testing process into a combinatorial optimization problem. Existing testing methods designed for DNN-based software can be applied to LLM-based software by AORTA, but their effectiveness is limited. To address this, we propose a novel testing method for LLM-based software within AORTA called Adaptive Beam Search. ABS is tailored for the expansive feature space of LLMs and improves testing effectiveness through an adaptive beam width and the capability for backtracking. We successfully embed 18 test methods in the designed framework AORTA and compared the test validity of ABS with three datasets and five threat models. ABS facilitates a more comprehensive and accurate robustness assessment before software deployment, with an average test success rate of 86.138%. Compared to the currently best-performing baseline PWWS, ABS significantly reduces the computational overhead by up to 3441.895 seconds per successful test case and decreases the number of queries by 218.762 times on average. Furthermore, test cases generated by ABS exhibit greater naturalness and transferability. © 2024, CC BY-NC-ND.

关键词： software testing

来源：评论

学校读者我要写书评

暂无评论

RaftFed: A Lightweight Federated Learning Framework for Vehicular Crowd Intelligence

arXiv

引用

arXiv 2023年

作者： Yang, Changan Chen, Yaxing Zhang, Yao Cui, Helei Yu, Zhiwen Guo, Bin Yan, Zheng Yang, Zijiang School of Software Northwestern Polytechnical University Xi’an China School of Computer Science Northwestern Polytechnical University Xi’an China School of Cyber Engineering Xidian University Xi’an China Turing Research Center for Interdisciplinary Information Science Xi'an Jiaotong University Xi’an China

Vehicular crowd intelligence (VCI) is an emerging research field. Facilitated by state-of-the-art vehicular ad-hoc networks and artificial intelligence, various VCI applications come to place, e.g., collaborative sensing, positioning, and mapping. The collaborative property of VCI applications generally requires data to be shared among participants, thus forming network-wide intelligence. How to fulfill this process without compromising data privacy remains a challenging issue. Although federated learning (FL) is a promising tool to solve the problem, adapting conventional FL frameworks to VCI is nontrivial. First, the centralized model aggregation is unreliable in VCI because of the existence of stragglers with unfavorable channel conditions. Second, existing FL schemes are vulnerable to Non-IID data, which is intensified by the data heterogeneity in VCI. This paper proposes a novel federated learning framework called RaftFed to facilitate privacy-preserving VCI. The experimental results show that RaftFed performs better than baselines regarding communication overhead, model accuracy, and model convergence. Copyright © 2023, The Authors. All rights reserved.

关键词： Vehicular ad hoc networks

来源：评论

学校读者我要写书评

暂无评论

WMAJL: Watcher-Mediated Attention Joint Learning Model for Multimodal Relation Extraction

WMAJL: Watcher-Mediated Attention Joint Learning Model for M...

引用

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

作者： Yunrui Dong Guiduo Duan Tianxi Huang Yunhao Li School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China Laboratory of Intelligent Collaborative Computing University of Electronic Science and Technology of China Chengdu China Trusted Cloud Computing and Big Data Key Laboratory of Sichuan Province Chengdu China Chengdu Textile College College of Humanities and General Education Chengdu China School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In the domain of Multimodal Relation Extraction (MRE), we present the $\color{Red}{\text{W}}$atcher-$\color{Red}{\text{M}}$ediated $\color{Red}{\text{A}}$ttention $\color{Red}{\text{J}}$oint $\color{Red}{\text{L}}$earning Model ($\color{Red}{\text{WMAJL}}$), a novel approach addressing the challenges of modality alignment noise, cross-modal fusion disparity, preservation of textual relative position information, and the distinctiveness of classification labels. WMAJL employs an integrative framework leveraging contrastive learning and variational autoencoder constraints to mitigate modality alignment noise by prioritizing relevant semantic data and effectively reducing extraneous noise that does not contribute to the task. The model’s innovative architecture includes a mediator watcher, which facilitates enhanced cross-modal fusion by enabling nuanced information exchange between textual and visual modalities while preserving the unique characteristics of each modality. Additionally, the design of auxiliary tasks, such as Named Entity Recognition (NER), and output supervision constructs loss functions that preserve relative position information, ensuring a precise depiction of entity relationships throughout the multilayer encoding processes. A key differentiator of WMAJL is its label-centric self-information loss technique, inspired by InfoNCE, which trains the model to cluster similar relation labels in semantically coherent areas, thereby optimizing classification label uniqueness by discerning subtle differences among relation types. The synergistic application of these strategies has led to a significant enhancement of WMAJL’s performance, as evidenced by its state-of-the-art F1 score of $\color{Red}{84.93\%}$ on the MNRE dataset. This achievement surpasses existing benchmarks and sets a new standard for multimodal knowledge extraction, underscoring WMAJL’s potential to revolutionize the MRE landscape.

关键词： Resistance Visualization Technological innovation Noise Autoencoders Semantics Contrastive learning Nonhomogeneous media Speech processing Standards

来源：评论

学校读者我要写书评

暂无评论

Split and Merge: Component Based Segmentation Network for Text Detection 1

引用

2nd International Conference on Pattern Recognition and Artificial Intelligence, ICPRAI 2020

作者： Gao, Pan Wan, Qi Shen, Linlin Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University Shenzhen China

ISBN: (数字)9783030598303

ISBN: (纸本)9783030598297

This paper presents a novel component-based detector to locate scene texts with arbitrary orientations, shapes and lengths. Our approach detects text by predicting four components like text region (TR), text skeleton (TS), text sub-region (TSR) and text connector (TC). TR and TS can well separate adjacent text instance. TSR are merged by TC to form a complete text instance. Experimental results show that the proposed approach outperforms state-of-the-art methods on two curved text datasets, i.e. 82.42% and 82.63% F-measures were achieved for the Total-Text and CTW1500, respectively. Our approach also achieves competitive performance on multi-oriented dataset, i.e. 85.86% f-measure for the ICDAR2015 was achieved. © 2020, Springer Nature Switzerland AG.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

TAD-Bench: A Comprehensive Benchmark for Embedding-Based Text Anomaly Detection

arXiv

引用

arXiv 2025年

作者： Cao, Yang Yang, Sikun Li, Chen Xiang, Haolong Qi, Lianyong Liu, Bo Li, Rongsheng Liu, Ming School of Computing and Information Technology Great Bay University China Great Bay Institute for Advanced Study Great Bay University China Graduate School of Informatics Nagoya University Japan School of Software Nanjing University of Information Science and Technology China College of Computer Science and Technology China University of Petroleum [East China China College of Cyberspace Security Zhengzhou University China School of Computer Harbin Engineering University China School of IT Deakin University Australia

Text anomaly detection is crucial for identifying spam, misinformation, and offensive language in natural language processing tasks. Despite the growing adoption of embedding-based methods, their effectiveness and generalizability across diverse application scenarios remain under-explored. To address this, we present TAD-Bench, a comprehensive benchmark designed to systematically evaluate embedding-based approaches for text anomaly detection. TAD-Bench integrates multiple datasets spanning different domains, combining state-of-the-art embeddings from large language models with a variety of anomaly detection algorithms. Through extensive experiments, we analyze the interplay between embeddings and detection methods, uncovering their strengths, weaknesses, and applicability to different tasks. These findings offer new perspectives on building more robust, efficient, and generalizable anomaly detection systems for real-world applications. Copyright © 2025, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

DISTDET: a cost-effective distributed cyber threat detection system 23

DISTDET: a cost-effective distributed cyber threat detection...

引用

Proceedings of the 32nd USENIX Conference on Security Symposium

作者： Feng Dong Liu Wang Xu Nie Fei Shao Haoyu Wang Ding Li Xiapu Luo Xusheng Xiao School of Cyber Science and Engineering Huazhong University of Science and Technology and Sangfor Technologies Inc Beijing University of Posts and Telecommunications Case Western Reserve University School of Cyber Science and Engineering Huazhong University of Science and Technology Key Laboratory of High-Confidence Software Technologies (MOE) School of Computer Science Peking University The Hong Kong Polytechnic University Arizona State University

ISBN: (纸本)9781939133373

Building provenance graph that considers causal relationships among software behaviors can better provide contextual information of cyber attacks, especially for advanced attacks such as Advanced Persistent Threat (APT) attacks. Despite its promises in assisting attack investigation, existing approaches that use provenance graphs to perform attack detection suffer from two fundamental limitations. First, existing approaches adopt a centralized detection architecture that sends all system auditing logs to the server for processing, incurring intolerable costs of data transmission, data storage, and computation. Second, they adopt either rule-based techniques that cannot detect unknown threats, or anomaly-detection techniques that produce numerous false alarms, failing to achieve a balance of precision and recall in APT detection. To address these fundamental challenges, we propose DISTDET, a distributed detection system that detects APT attacks by (1) performing light weight detection based on the host model built in the client side, (2) filtering false alarms based on the semantics of the alarm proprieties, and (3) deriving global models to complement the local bias of the host models. Our experiments on a large-scale industrial environment (1,130 hosts, 14 days, ∼1.6 billion events) and the DARPA TC dataset show that DISTDET is as effective as sate-of-the-art techniques in detecting attacks, while dramatically reducing network bandwidth from 11.28Mb/s to 17.08Kb/S (676.5× reduction), memory usages from 364MB to 5.523MB (66× reduction), and storage from 1.47GB to 130.34MB (11.6× reduction). By the time of this writing, DISTDET has been deployed to 50+ industry customers with 22,000+ hosts for more than 6 months, and identified over 900 real-world attacks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Detecting Memory-Related Bugs by Tracking Heap Memory Management of C++ Smart Pointers 36

Detecting Memory-Related Bugs by Tracking Heap Memory Manage...

引用

36th IEEE/ACM International Conference on Automated software engineering, ASE 2021

作者： Ma, Xutong Yan, Jiwei Wang, Wei Yan, Jun Zhang, Jian Qiu, Zongyan Chinese Academy of Sciences State Key Laboratory of Computer Science Institute of Software China Technology Center of Software Engineering Institute of Software Chinese Academy of Sciences China University of Chinese Academy of Sciences China Peking University School of Mathematical Sciences China

ISBN: (纸本)9781665403375

The smart pointer mechanism, which is improved in the continuous versions of the C++ standards over the last decade, is designed to prevent memory-leak bugs by automatically deallocating the managed memory blocks. However, not all kinds of memory errors can be immunized by adopting this mechanism. For example, dereferencing a null smart pointer will lead to a software failure. Due to the lack of specialized support for smart pointers, the off-the-shelf C++ static analyzers cannot effectively reveal these *** this paper, we propose a static approach to detecting memory-related bugs by tracking the heap memory management of smart pointers. The behaviors of smart pointers are modeled during their lifetime to trace the state transitions of managed memory blocks. and the specially designed checkers are used to check the state changes according to five collected bug patterns. To evaluate the effectiveness of our approach, we implement it on the top of the Clang Static Analyzer. A set of handmade code snippets, as well as nine popular open-source C++ projects, are used to compare our tool against four other analyzers. The results show that our approach can successfully discover nearly all the built-in bugs. and 442 out of 648 reports generated from the open-source projects are true positives after manual reviewing, where the bugs of dereferencing null smart pointers are most frequently reported. To further confirm our reports, we design patches for Aria2, Restbed, MySQL and LLVM, in which seven pull requests covering 76 bug reports have been merged by the developers up to now. The results indicate that pointers should always be carefully used even after migrated to smart pointers and static analysis upon specialized models can effectively detect such bugs. © 2021 IEEE.

关键词： Static analysis

来源：评论

学校读者我要写书评

暂无评论

Composite Virtual Spatial Modulation and Its Diversity Enhancement

Composite Virtual Spatial Modulation and Its Diversity Enhan...

引用

IEEE International Conference on Communications Workshops, ICC

作者： Zahid Iqbal Fei Ji Jun Li Qiang Li Adil Nawaz Xue-Qin Jiang Jia Hou School of Electronic and Information Engineering South China University of Technology Guangzhou China School of Electronics and Communication Engineering Guangzhou University Guangzhou China College of Information Science and Technology Jinan University Guangzhou China College of Computer Science and Software Engineering Shenzhen University Shenzhen China School of Information Science and Technology Donghua University Shanghai China School of Electronic and Information Engineering Soochow University Suzhou China

Virtual spatial modulation (VSM) has arisen as a promising technique for improving the energy efficiency (EE) of multiple-input multiple-output (MIMO) systems. However, due to utilizing only a subset of virtual parallel channels, the spectral efficiency (SE) of VSM is degraded compared to the classical MIMO system. To overcome this drawback, we propose an index and composition VSM (IC-VSM) technique to enhance the SE of the VSM scheme. In IC-VSM, the indices of active/inactive codeword of virtual parallel channels and compositions of an integer number are utilized to encode information, where the information bits compose of the channel activation patterns (CAPs), modulation symbols, as well as the energy allocation patterns (EAPs). We further propose a diversity-enhanced IC-VSM scheme through coordinate interleaving (CI). The average bit error probability (ABEP), asymptotic ABEP, and the diversity order are also derived to evaluate the performance of IC-VSM. Simulation results show that IC-VSM scheme outperforms the state-of-the-art SM schemes, i.e., VSM, particularly at high signal-to-noise ratio (SNR), and corroborate the validity of ABEP theoretical analysis.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Near-Field Full Dimensional Beam Codebook Design for XL-MIMO Communications

Near-Field Full Dimensional Beam Codebook Design for XL-MIMO...

引用

IEEE Conference on Global Communications (GLOBECOM)

作者： Wei Huang Cuiling Li Yong Zeng Caihong Kai Shiwen He School of Comnuting Science and Information Hefei University of Technology Hefei China National Mobile Communications Research Laboratory Southeast University Nanjing China School of Computer Science and Engineering Central South University Changsha China School of Software Xinjiang University Urumqi China

Extremely large-scale multiple-input multiple-output (XL-MIMO) communication system with extremely large-scale antenna arrays can achieve ultra-high spectral efficiency. However, the conventional far-field beam codebooks may be mismatched with the near-field spherical-wavefront channel caused by large array aperture, which results in severe performance loss. To address this problem, we develop a criterion of code book design to maximize the worst-case beam gain within the beam coverage. Then, a closed-form expression of the near-field full dimensional (FD) codebook with non-orthogonal structure is derived, which can realize the spatial oversampling regardless of the number of antennas at the transceiver. Simulation results show that our proposed non-orthogonal codebook can potentially improve the accuracy of near-field beam training, compared with existing codebooks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Analytic Correlation Penalty with Variable Window in Multi-task Learning Disease Progression Model

Analytic Correlation Penalty with Variable Window in Multi-t...

引用

International Conference on Mobile Ad-hoc and Sensor Networks, MSN

作者： Xiangchao Chang Menghui Zhou Fengtao Nan Yun Yang Po Yang School of Software Yunnan University Kunming China School of Information Science and Engineering Yunnan University Kunming China Department of Computer Science University of Sheffield Sheffield UK

Alzheimer's Disease (AD) is the most common reason of dementia that causes serious problems in patients' congnitive functions. Multi-task learning (MTL) has performed well in studies of longitudinal processes in Alzheimer's disease for revealing the progression of AD. Combined with prior knowl-edges in disease progression or medical science, regularization MTL framework could introduce empirical constraints more flexibly. Meanwhile, it brings higher cost during optimization. While it shown that most of formulations could not define the disease progression precisely. Existing regression methods with temporal smoothness method eliminated abnormal fluctuation of cognitive scores, and neglected the sophisticated progression in disease. In this article, we proposed an analytic method to define the progression of AD, and a flexible bandwidth method to encourage the points of disease time sequence temporal smoothness in an appropriate way. To solve three non-smooth penalties in our method, we proposed an optimization method combined accelerated gradient descent (AGD) and alternating direction method of multipliers (ADMM).

关键词： Analytical models Statistical analysis Optimization methods Information sharing Bandwidth Predictive models Multitasking

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：