检索结果-南通市图书馆

维普期刊数据库

SwiftTheft: A Time-Efficient Model Extraction Attack Framework Against Cloud-Based Deep Neural Networks

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第1期33卷 90-100页

作者： Wenbin YANG Xueluan GONG Yanjiao CHEN Qian WANG Jianshuo DONG school of Cyber science and Engineering Wuhan University school of computer science Wuhan University College of electrical Engineering Zhejiang University

With the rise of artificial intelligence and cloud computing, machine-learning-as-a-service platforms,such as Google, Amazon, and IBM, have emerged to provide sophisticated tasks for cloud applications. These proprietary models are vulnerable to model extraction attacks due to their commercial value. In this paper, we propose a time-efficient model extraction attack framework called Swift Theft that aims to steal the functionality of cloud-based deep neural network models. We distinguish Swift Theft from the existing works with a novel distribution estimation algorithm and reference model settings, finding the most informative query samples without querying the victim model. The selected query samples can be applied to various cloud models with a one-time selection. We evaluate our proposed method through extensive experiments on three victim models and six datasets, with up to 16 models for each dataset. Compared to the existing attacks, Swift Theft increases agreement(i.e., similarity) by 8% while consuming 98% less selecting time.

关键词： Performance evaluation Cloud computing Analytical models Computational modeling Estimation Artificial neural networks Artificial intelligence

Numerical simulation for the initial state of avalanche in polydisperse particle systems

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2024年第2期33卷 405-412页

作者：韩韧李亭迟志鹏杨晖李然 school of Optical-electrical and computer Engineering University of Shanghai for Science and TechnologyShanghai 200093China

Numerical simulation is employed to investigate the initial state of avalanche in polydisperse particle *** and propagation processes are illustrated for pentadisperse and triadisperse particle systems,*** these processes,particles involved in the avalanche grow slowly in the early stage and explosively in the later stage,which is clearly different from the continuous and steady growth trend in the monodisperse *** examining the avalanche propagation,the number growth of particles involved in the avalanche and the slope of the number growth,the initial state can be divided into three stages:T1(nucleation stage),T2(propagation stage),T3(overall avalanche stage).We focus on the characteristics of the avalanche in the T2 stage,and find that propagation distances increase almost linearly in both axial and radial directions in polydisperse *** also consider the distribution characteristics of the average coordination number and average velocity for the moving *** results support that the polydisperse particle systems are more stable in the T2 stage.

关键词： avalanche initial state polydisperse particle systems propagation

维普期刊数据库

Metarelation2vec:A Metapath-Free Scalable Representation Learning Model for Heterogeneous Networks

在线全文

学校读者我要写书评

暂无评论

Tsinghua science and Technology 2024年第2期29卷 553-575页

作者： Lei Chen Yuan Li Yong Lei Xingye Deng school of Information and electrical Engineering Hunan University of Science and TechnologyXiangtan 411201China school of computer science and Engineering Hunan University of Science and TechnologyXiangtan 411201China

Metapaths with specific complex semantics are critical to learning diverse semantic and structural information of heterogeneous networks(HNs)for most of the existing representation learning ***,any metapaths consisting of multiple,simple metarelations must be driven by domain *** sensitive,expensive,and limited metapaths severely reduce the flexibility and scalability of the existing models.A metapath-free,scalable representation learning model,called Metarelation2vec,is proposed for HNs with biased joint learning of all metarelations in a bid to address this ***,a metarelation-aware,biased walk strategy is first designed to obtain better training samples by using autogenerating cooperation probabilities for all metarelations rather than using expert-given ***,grouped nodes by the type,a common and shallow skip-gram model is used to separately learn structural proximity for each node ***,grouped links by the type,a novel and shallow model is used to separately learn the semantic proximity for each link ***,supervised by the cooperation probabilities of all meta-words,the biased training samples are thrown into the shallow models to jointly learn the structural and semantic information in the HNs,ensuring the accuracy and scalability of the *** experimental results on three tasks and four open datasets demonstrate the advantages of our proposed model.

关键词： metarelation random walk heterogeneous network metapath representation learning

BAD-FM:Backdoor Attacks Against Factorization-Machine Based Neural Network for Tabular Data Prediction

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第4期33卷 1077-1092页

作者： Lingshuo MENG Xueluan GONG Yanjiao CHEN College of electrical Engineering Zhejiang University school of computer science Wuhan University

Backdoor attacks pose great threats to deep neural network models. All existing backdoor attacks are designed for unstructured data(image, voice, and text), but not structured tabular data, which has wide real-world applications, e.g., recommendation systems, fraud detection, and click-through rate prediction. To bridge this research gap, we make the first attempt to design a backdoor attack framework, named BAD-FM, for tabular data prediction models. Unlike images or voice samples composed of homogeneous pixels or signals with continuous values, tabular data samples contain well-defined heterogeneous fields that are usually sparse and discrete. Tabular data prediction models do not solely rely on deep networks but combine shallow components(e.g., factorization machine, FM) with deep components to capture sophisticated feature interactions among fields. To tailor the backdoor attack framework to tabular data models, we carefully design field selection and trigger formation algorithms to intensify the influence of the trigger on the backdoored model. We evaluate BAD-FM with extensive experiments on four datasets, i.e.,HUAWEI, Criteo, Avazu, and KDD. The results show that BAD-FM can achieve an attack success rate as high as 100%at a poisoning ratio of 0.001%, outperforming baselines adapted from existing backdoor attacks against unstructured data models. As tabular data prediction models are widely adopted in finance and commerce, our work may raise alarms on the potential risks of these models and spur future research on defenses.

关键词： Adaptation models Systematics Frequency modulation Finance Predictive models Prediction algorithms Data models

Understanding Discrepancy of Power System Dynamic Security Assessment with Unknown Faults: A Reliable Transfer Learning-based Method

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CSEE Journal of Power and Energy Systems 2024年第1期10卷 427-431页

作者： Chao Ren Han Yu Yan Xu Zhao Yang Dong school of computer science and Engineering Nanyang Technological UniversitySingapore school of electrical and Electronic Engineering Nanyang Technological UniversitySingapore

This letter proposes a reliable transfer learning(RTL)method for pre-fault dynamic security assessment(DSA)in power systems to improve DSA performance in the presence of potentially related unknown *** takes individual discrepancies into consideration and can handle unknown faults with incomplete *** experiment results demonstrate high DSA accuracy and computational efficiency of the proposed RTL *** analysis shows RTL can guarantee system performance.

关键词： Adversarial training dynamic security assessment maximum classifier discrepancy missing data transfer learning

Post-layout simulation driven analog circuit sizing

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第4期67卷 289-300页

作者： Xiaohan GAO Haoyi ZHANG Siyuan YE Mingjie LIU David Z.PAN Linxiao SHEN Runsheng WANG Yibo LIN Ru HUANG school of Integrated Circuits Peking University school of computer science Peking University Department of electrical and computer Engineering University of Texas at Austin Institute of Electronic Design Automation Peking University

Post-layout simulation provides accurate guidance for analog circuit design, but post-layout performance is hard to be directly optimized at early design stages. Prior work on analog circuit sizing often utilizes pre-layout simulation results as the optimization objective. In this work, we propose a post-layoutsimulation-driven(post-simulation-driven for short) analog circuit sizing framework that directly optimizes the post-layout simulation performance. The framework integrates automated layout generation into the optimization loop of transistor sizing and leverages a coupled Bayesian optimization algorithm to search for the best post-simulation performance. Experimental results demonstrate that our framework can achieve over 20% better post-layout performance in competitive time than manual design and the method that only considers pre-layout optimization.

关键词： analog EDA transistor sizing Bayesian optimization post-layout simulation

What can we learn from quality assurance badges in open-source software?

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第4期67卷 146-163页

作者： Feng LI Yiling LOU Xin TAN Zhenpeng CHEN Jinhao DONG Yang LI Xuanzhi WANG Dan HAO Lu ZHANG MoE Key Lab of HCST School of Computer SciencePeking University Department of computer science Purdue University school of computer science & Engineering Beihang University(BUAA) Department of computer science University College London(UCL)

In the development of open-source software(OSS), many developers use badges to give an overview of the software and share some key features/metrics conveniently. Among various badges, quality assurance(QA) badges make up a large proportion and are the most prevalent because QA is of vital importance in software development, and ineffective QA may lead to anomalies or defects. In this paper, we focus on QA badges in open-source projects, which present quality assurance information directly and instantly,and aim to produce some interesting findings and provide practical implications. We collect and analyze 100000 projects written in popular programming languages from GitHub and conduct a comprehensive empirical study both inside and outside QA badges. Inside QA badges, we build a category classification for all QA badges based on the properties they focus on, which shows the types of QA badges developers use. Then,we analyze the frequency of the properties that QA badges focus on, and property combinations, too, which present their use status. We find that QA badges focus on various properties while developers give different preferences to different properties. The use status also differs between different programming languages. For example, projects written in C focus on Security to a great extent. Our findings also provide implications for developers and badge providers. Outside QA badges, we conduct a correlation analysis between QA badges and some software metrics that have potential relationships with code quality, contribution quality, and popularity. We find that QA badges have statistically significant correlations with various software metrics.

关键词： quality assurance badge open-source software code quality empirical study

SDCC: software-defined collective communication for distributed training

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第9期67卷 105-125页

作者： Xin JIN Zhen ZHANG Yunshan JIA Yun MA Xuanzhe LIU school of computer science Peking University Department of computer science and Technology Johns Hopkins University

Communication is crucial to the performance of distributed training. Today's solutions tightly couple the control and data planes and lack flexibility, generality, and performance. In this study, we present SDCC, a software-defined collective communication framework for distributed training. SDCC is based on the principle of modern systems design to effectively decouple the control plane from the data *** abstracts the operations for collective communication in distributed training with dataflow operations and unifies computing and communication with a single dataflow graph. The abstraction, together with the unification, is powerful: it enables users to easily express new and existing collective communication algorithms and optimizations, simplifies the integration with different computing engines(e.g., Py Torch and Tensor Flow) and network transports(e.g., Linux TCP and kernel bypass), and allows the system to improve performance by exploiting parallelism exposed by the dataflow graph. We further demonstrate the benefits of SDCC in four use cases.

关键词： machine learning systems distributed training deep learning collective communication software-defined networking

UAMP-Based Delay-Doppler Channel Estimation for OTFS Systems

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

China Communications 2024年第10期 1-15页

作者： Li Zhongjie Yuan Weijie Guo Qinghua Wu Nan Zhang Ji Department of Electronic and electrical Engineering Southern University of Science and Techonology school of electrical Computer and Telecommunications Engineering University of Wollongong school of Integrated Circuits and Electronics Beijing Institute of Technology school of Mathematics and Statistics Henan University of Science and Technology

Orthogonal time frequency space(OTFS)technique, which modulates data symbols in the delayDoppler(DD) domain, presents a potential solution for supporting reliable information transmission in highmobility vehicular networks. In this paper, we study the issues of DD channel estimation for OTFS in the presence of fractional Doppler. We first propose a channel estimation algorithm with both low complexity and high accuracy based on the unitary approximate message passing(UAMP), which exploits the structured sparsity of the effective DD domain channel using hidden Markov model(HMM). The empirical state evolution(SE) analysis is then leveraged to predict the performance of our proposed algorithm. To refine the hyperparameters in the proposed algorithm,we derive the update criterion for the hyperparameters through the expectation-maximization(EM) algorithm. Finally, Our simulation results demonstrate that our proposed algorithm can achieve a significant gain over various baseline schemes.

关键词： channel estimation hidden Markov model (HMM) orthogonal time frequency space (OTFS) unitary approximate message passing (UAMP)