检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Peng, Minlong Xiong, Zidi Nguyen, Quang H. Sun, Mingming Doan, Khoa D. Li, Ping Cognitive Computing Lab Baidu Research No.10 Xibeiwang East Road Beijing100193 China 10900 NE 8th St. Bellevue Washington98004 United States College of Engineering and Computer Science VinUniversity Viet Nam

Backdoor attacks aim to manipulate the victim model into producing specific outputs on any input injected with pre-designed triggers. Existing dirty-label backdoor attacks, despite showing high attack efficiency, suffer from the overconfidence problem: the victim models will behave peculiarly when the backdoor is implanted. Built upon this characteristic, existing defense algorithms are highly effective in mitigating the risks of these attacks. This work proposes a novel method to overcome this over-confidence problem in the existing attacks and accordingly, increase their stealthiness against the existing defense algorithms relying on this over-confidence characteristic. The rationale of our method is to reduce the backdoor effect in such a way that the victim model will only predict the poisoned input as the target class with a probability slightly greater than those of the other classes. This is achieved by re-designing the label-changing strategy of dirty-label backdoor attacks: the label of a poisoned input x will be changed to the target class with an instance-specific probability of pt(x), instead of 100% as in previous methods. We empirically show that our method considerably improves the stealthiness of several representative backdoor attacks against the defense algorithms on four benchmark datasets in the backdoor domain. The results encourage backdoor researchers to develop defensive countermeasures to mitigate this type of attack. © 2022, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

LAS-AT: Adversarial Training with Learnable Attack Strategy

arXiv

引用

arXiv 2022年

作者： Jia, Xiaojun Zhang, Yong Wu, Baoyuan Ma, Ke Wang, Jue Cao, Xiaochun Institute of Information Engineering Chinese Academy of Sciences Beijing China School of Cyberspace Security University of Chinese Academy of Sciences Beijing China Tencent AI Lab Shenzhen China School of Data Science The Chinese University of Hong Kong Shenzhen China Secure Computing Lab of Big Data Shenzhen Research Institute of Big Data Shenzhen China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

Adversarial training (AT) is always formulated as a minimax problem, of which the performance depends on the inner optimization that involves the generation of adversarial examples (AEs). Most previous methods adopt Projected Gradient Decent (PGD) with manually specifying attack parameters for AE generation. A combination of the attack parameters can be referred to as an attack strategy. Several works have revealed that using a fixed attack strategy to generate AEs during the whole training phase limits the model robustness and propose to exploit different attack strategies at different training stages to improve robustness. But those multi-stage hand-crafted attack strategies need much domain expertise, and the robustness improvement is limited. In this paper, we propose a novel framework for adversarial training by introducing the concept of "learnable attack strategy", dubbed LAS-AT, which learns to automatically produce attack strategies to improve the model robustness. Our framework is composed of a target network that uses AEs for training to improve robustness, and a strategy network that produces attack strategies to control the AE generation. Experimental evaluations on three benchmark databases demonstrate the superiority of the proposed method. The code is released at https://***/jiaxiaojunQAQ/LAS-AT. Copyright © 2022, The Authors. All rights reserved.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Single-cell technologies: current and near future

引用

science China(Life sciences) 2025年第1期68卷 1-4页

作者： Chenfei Wang Qi Liu Xiaohui Fan Tieliu Shi KeyLaboratory of Spine and Spinal CordInjury Repairand Regeneration Ministry of EducationOrthopedics DepartmentTongji HospitalBioinformatics DepartmentSchoolofLifeSciencesandTechnologyTongji UniversityShanghai 200082China Frontier Science Center for Stem Cells School of Life Sciences and TechnologyTongji UniversityShanghai 200092China Translational Medical Center for Stem Cell Therapy and Institute for Regenerative Medicine Shanghai East HospitalBioinformatics DepartmentSchool of Life Sciences and TechnologyTongji UniversityShanghai 200082China Research Institute of Intelligent Computing Zhejiang LabHangzhou 311121China Pharmaceutical Informatics Institute College of Pharmaceutical SciencesZhejiang UniversityHangzhou 310058China National Key Laboratory of Chinese Medicine Modernization Innovation Center of Yangtze River DeltaZhejiang UniversityJiaxing 314103China Center for Bioinformatics and Computational Biology Shanghai Key Laboratory of Regulatory Biologythe Institute of Biomedical Sciences and School of Life SciencesEast China Normal UniversityShanghai 200241China Key Laboratory of Advanced Theory and Application in Statistics and Data Science(MOE) School of StatisticsEast China Normal UniversityShanghai 200062China

Single-cell technologies enable the indepth exploration of multiple biological hierarchies at the scale of individual cells,which have deepened our knowledge of cellular diversity,tissue organization,and overall organism function(Sun et al.,2024).In 2009,single-cell RNA sequencing(scRNA-seq)was first developed as a powerful tool to dissect gene expression and uncover transcriptional *** the rapid advancement of technology,single-cell approaches have expanded to encompass other omics,such as genomics,epigenomics,proteomics,and ***,multimodal technologies including spatially resolved transcriptomics and clustered regularly interspaced short palindromic repeats(CRISPR)screening at the single-cell level further advanced our ability to comprehensively understand cellular ***-cell methods provide deeper insights into disease mechanisms across various conditions,including cancer,developmental disorders,and aging-associated *** this issue,we focus on topics related to single-cell methods,covering the following four aspects:(i)single-cell multi-omic sequencing technology;(ii)single-cell spatial technology;(iii)singlecell CRISPR screening technology;(iv)applications of single-cell technology.

关键词： expanded enable insight

来源：评论

学校读者我要写书评

暂无评论

CPSketch: A ‘couple’ sketch-based heavy flow detection method

引用

Journal of Network and Computer Applications 2025年 242卷

作者： Renpin Yao Yang Cao Yunhe Cui Yi Chen Chun Guo Guowei Shen Engineering Research Center of Text Computing & Cognitive Intelligence Guizhou University China College of Computer Science and Technology Guizhou University China State Key Laboratory of Public Big Data Guizhou University China

One of the core challenges in network measurement for large-scale networks is the accurate and efficient identification of heavy flows. This task has grown increasingly complex due to limited memory resources and the dynamic nature of network traffic patterns. Most current heavy flow detection methods fail to fully consider the mutual collisions between large flows and their sparse distribution, which are caused by the combined effects of hash collisions and the heavy-tailed distribution of network traffic, resulting in an inefficient use of limited memory resources. Additionally, existing approaches uniformly treat all flows. As a result, when the network contains many small flows, heavy flows are frequently and erroneously replaced. To address these issues, we propose CPSketch, a novel method that combines sketch with ‘couple’ buckets to enhance heavy flow detection accuracy. CPSketch optimizes memory utilization by leveraging cold buckets to store more heavy flows. Specifically, it accurately estimates flow sizes by monitoring the number of matched packets for candidate flows in real-time. For small flows, CPSketch rapidly identifies and evicts them based on flow ratios. For other flows, CPSketch calculates replacement probabilities using multidimensional statistical information. To protect heavy flows that have been incorrectly evicted, CPSketch employs a global hash function to provide additional storage opportunities and dynamically constructs couple buckets for these heavy flows. Moreover, CPSketch extends the selection range for minimum flows by leveraging couple buckets in the flow replacement policy, effectively mitigating the impact of hash limitations and improving memory efficiency. Experimental results from trace-driven simulations and Open vSwitch (OVS) tests demonstrate that CPSketch outperforms existing methods in accuracy, even under constrained memory and high traffic loads. It achieves improvements of up to 14.74 % in F1-score while reducing the

关键词：

来源：评论

学校读者我要写书评

暂无评论

Which Pixel to Annotate: a label-Efficient Nuclei Segmentation Framework

arXiv

引用

arXiv 2022年

作者： Lou, Wei Li, Haofeng Li, Guanbin Han, Xiaoguang Wan, Xiang Shenzhen Research Institute of Big Data Guangdong Provincial Key Laboratory of Big Data Computing The Chinese University of Hong Kong at Shenzhen Shenzhen518172 China The School of Computer Science and Engineering Sun Yat-sen University Guangzhou510006 China Pazhou Lab Guangzhou510330 China

Recently deep neural networks, which require a large amount of annotated samples, have been widely applied in nuclei instance segmentation of H&E stained pathology images. However, it is inefficient and unnecessary to label all pixels for a dataset of nuclei images which usually contain similar and redundant patterns. Although unsupervised and semi-supervised learning methods have been studied for nuclei segmentation, very few works have delved into the selective labeling of samples to reduce the workload of annotation. Thus, in this paper, we propose a novel full nuclei segmentation framework that chooses only a few image patches to be annotated, augments the training set from the selected samples, and achieves nuclei segmentation in a semi-supervised manner. In the proposed framework, we first develop a novel consistency-based patch selection method to determine which image patches are the most beneficial to the training. Then we introduce a conditional single-image GAN with a component-wise discriminator, to synthesize more training samples. Lastly, our proposed framework trains an existing segmentation model with the above augmented samples. The experimental results show that our proposed method could obtain the same-level performance as a fully-supervised baseline by annotating less than 5% pixels on some benchmarks. Copyright © 2022, The Authors. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Classification of Breast Thermal Images into Healthy/Cancer Group Using Pre-Trained Deep Learning Schemes

Classification of Breast Thermal Images into Healthy/Cancer ...

引用

2022 International Conference on Machine Learning and data Engineering, ICMLDE 2022

作者： Kadry, Seifedine Crespo, Rubén González Herrera-Viedma, Enrique Krishnamoorthy, Sujatha Rajinikanth, Venkatesan Faculty of Applied Computing and Technology Noroff University College Kristiansand94612 Norway Computer Science Department School of Engineering and Technology Universidad Internacional de la Rioja Andalusia26006 Spain Research Institute in Data Science and Computational Intelligence University of Granada Granada Spain Zhejiang Bioinformatics International Science and Technology Cooperation Center Wenzhou-Kean University Zhejiang Province China Wenzhou Municipal Key Lab of Applied Biomedical and Biopharmaceutical Informatics Wenzhou-Kean University Zhejiang Province China Department of Computer Science and Engineering Saveetha School of Engineering SIMATS Tamil Nadu Chennai602105 India

In the women's community, Breast Cancer (BC) is a severe disease. The World Health Organization reported in 2020 that 2.26 million deaths occur due to BC. BC is curable if detected early. Since thermal imaging is non-invasive and supports disease detection, it is commonly used in clinics. Compared to other methods, it keeps BC early and accurate. The proposed work aims to evaluate the performance of the Pretrained Deep-Learning Methods (PDLM) in detecting BC using the thermal images collected from the benchmark dataset. It includes the following stages: primary image processing, deep feature mining, handcrafted feature mining, feature optimization using Firefly-Algorithm (FA), classification and validation. Visual lab thermal images were used in the study. The investigational outcome of this study authenticates that the VGG16, along with the DT, provides better detection accuracy (95.5%) compared to other classifiers used in this study. To justify the significance of the implemented technique, the proposed work not only improved accuracy, but also improved precision, sensitivity, specificity, and F1-Scores. © 2023 The Authors. Published by Elsevier B.V.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

INMO: A Model-Agnostic and Scalable Module for Inductive Collaborative Filtering

arXiv

引用

arXiv 2021年

作者： Wu, Yunfan Cao, Qi Shen, Huawei Tao, Shuchang Cheng, Xueqi Data Intelligence System Research Center Institute of Computing Technology CAS University of Chinese Academy of Sciences Beijing China Data Intelligence System Research Center Institute of Computing Technology CAS China CAS Key Lab of Network Data Science and Technology Institute of Computing Technology CAS University of Chinese Academy of Sciences Beijing China

Collaborative filtering is one of the most common scenarios and popular research topics in recommender systems. Among existing methods, latent factor models, i.e., learning a specific embedding for each user/item by reconstructing the observed interaction matrix, have shown excellent performances. However, such user-specific and item-specific embeddings are intrinsically transductive, making it difficult to deal with new users and new items unseen during training. Besides, the number of model parameters heavily depends on the number of all users and items, restricting its scalability to real-world applications. To solve the above challenges, in this paper, we propose a novel model-agnostic and scalable Inductive Embedding Module for collaborative filtering, namely INMO. INMO generates the inductive embeddings for users (items) by characterizing their interactions with some template items (template users), instead of employing an embedding lookup table. Under the theoretical analysis, we further propose an effective indicator for the selection of template users/items. Our proposed INMO can be attached to existing latent factor models as a pre-module, inheriting the expressiveness of backbone models, while bringing the inductive ability and reducing model parameters. We validate the generality of INMO by attaching it to both Matrix Factorization (MF) and LightGCN, which are two representative latent factor models for collaborative filtering. Extensive experiments on three public benchmarks demonstrate the effectiveness and efficiency of INMO in both transductive and inductive recommendation scenarios. © 2021, CC BY.

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

JARVIS-Leaderboard:a large scale benchmark of materials design methods

引用

npj Computational Materials 2024年第1期10卷 2280-2296页

作者： Kamal Choudhary Daniel Wines Kangming Li Kevin F.Garrity Vishu Gupta Aldo H.Romero Jaron T.Krogel Kayahan Saritas Addis Fuhr Panchapakesan Ganesh Paul R.C.Kent Keqiang Yan Yuchao Lin Shuiwang Ji Ben Blaiszik Patrick Reiser Pascal Friederich Ankit Agrawal Pratyush Tiwary Eric Beyerle Peter Minch Trevor David Rhone Ichiro Takeuchi Robert B.Wexler Arun Mannodi-Kanakkithodi Elif Ertekin Avanish Mishra Nithin Mathew Mitchell Wood Andrew Dale Rohskopf Jason Hattrick-Simpers Shih-Han Wang Luke E.K.Achenie Hongliang Xin Maureen Williams Adam J.Biacchi Francesca Tavazza Material Measurement Laboratory National Institute of Standards and TechnologyGaithersburgMD20899USA Department of Materials Science and Engineering University of Toronto27 King’s College CirTorontoONCanada Department of Electrical and Computer Engineering Northwestern UniversityEvanstonIL 60208USA Lewis-Sigler Institute for Integrative Genomics Princeton UniversityPrincetonNJ 08544USA Ludwig Institute for Cancer Research Princeton UniversityPrincetonNJ 08544USA Department of Physics and Astronomy West Virginia UniversityMorgantownWV 26506USA Materials Science and Technology Division Oak Ridge National LaboratoryOak RidgeTN 37831USA Center for Nanophase Materials Science Oak Ridge National LaboratoryOak RidgeTN 37831USA Computational Sciences and Engineering Division Oak Ridge National LaboratoryOak RidgeTN 37831USA Department of Computer Science and Engineering Texas A&M UniversityCollege StationTX 77843USA Globus University of ChicagoChicagoIL 60637USA Data Science and Learning Division Argonne National LabLemontIL 60439USA Institute of Nanotechnology Karlsruhe Institute of TechnologyKaiserstraße 1276131 KarlsruheGermany Institute of Theoretical Informatics Karlsruhe Institute of TechnologyKaiserstraße 1276131 KarlsruheGermany Department of Chemistry and Biochemistry and Institute for Physical Science and Technology University of MarylandCollege ParkMD 20742USA Department of Physics Applied Physics and AstronomyRensselaer Polytechnic InstituteTroyNY12180USA Department of Materials Science and Engineering University of MarylandCollege ParkMD20742USA Department of Chemistry and Institute of Materials Science and Engineering Washington University in St.LouisSt.LouisMO63130USA School of Materials Engineering Purdue UniversityWest LafayetteIN47907USA Department of Mechanical Science and Engineering University of Illinois Urbana-ChampaignUrbanaIllinois 61801USA Materials Research Laboratory University of Illinois Urbana-ChampaignUrbanaIL 61801USA Theore

Lack of rigorous reproducibility and validation are significant hurdles for scientific development across many *** science,in particular,encompasses a variety of experimental and theoretical approaches that require careful *** efforts have been developed previously to mitigate these ***,a comprehensive comparison and benchmarking on an integrated platform with multiple data modalities with perfect and defect materials data is still *** work introduces JARVIS-Leaderboard,an open-source and community-driven platform that facilitates benchmarking and enhances *** platform allows users to set up benchmarks with customtasks and enables contributions in the form of dataset,code,and meta-data *** cover the following materials design categories:Artificial Intelligence(AI),Electronic Structure(ES).

关键词： rigorous perfect enable

来源：评论

学校读者我要写书评

暂无评论

Automatically derived stateful network functions including non-field attributes

Automatically derived stateful network functions including n...

引用

IEEE International Conference on Trust, Security and Privacy in computing and Communications (TrustCom)

作者： Bin Yuan Shengyao Sun Xianjun Deng Deqing Zou Haoyu Chen Shenghui Li Hai Jin School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Engineering Research Center on Big Data Security Shenzhen Research Institute Huazhong University of Science and Technology Shenzhen China School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China Cluster and Grid Computing Lab

ISBN: (纸本)9781665416597

The modern network consists of thousands of network devices from different suppliers that perform distinct code-pendent functions, such as routing, switching, modifying header fields, and access control across physical and virtual networks. Because of the network complexity, the network is prone to a wide range of errors, such as false-positive configuration, software errors, or unexpected interactions across protocols. These errors can lead to loops, sub-optimal routing, path leaks, black holes, and access control violations that make services unavailable, vulnerable to exploitation, or prone to attacks (e.g., DDoS attacks). To mitigate these problems, network operators deploy many different stateful network functions, like firewalls, NATs, load balancers, and intrusion-prevention boxes. They have become an important part of networks today, so it is critical to verify that these network functions are the same as expected deployments. All static network verification tools are meant to rigorously check network software or configuration for bugs before deployment. They usually use handwritten models or limited derivation models that are error-prone and ignore the fact that even the same type of network functions (from different vendors) still have different implementation details. In this paper, we propose a tool that can automatically synthesize more realistic and high-fidelity models that include stateful network functions with non-field attributes. We design an inferring algorithm, implement the transformation between data packages and symbolic packages, and obtain a finite state machine that can accurately express the actions of black-box network functions for a given configuration.

关键词： Access control Performance evaluation Privacy Protocols Computational modeling Switches Routing

来源：评论

学校读者我要写书评

暂无评论

基于污点和概率的逃逸恶意软件多路径探索

引用

Security and Safety 2023年第3期2卷 83-106页

作者：徐钫洲张网羌卫中金海 National Engineering Research Center for Big Data Technology and System Wuhan 430074China Services Computing Technology and System Lab Cluster and Grid Computing LabWuhan 430074China Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data SecurityWuhan 430074China School of Cyber Science and Engineering Huazhong University of Science and TechnologyWuhan 430074China School of Computer Science and Technology Huazhong University of Science and TechnologyWuhan 430074China Jinyinhu Laboratory Wuhan 430040China

Static analysis is often impeded by malware obfuscation techniques,such as encryption and packing,whereas dynamic analysis tends to be more resistant to obfuscation by leveraging concrete execution ***,malware can employ evasive techniques to detect the analysis environment and alter its behavior *** known evasive techniques can be explicitly dismantled,the challenge lies in generically dismantling evasions without full knowledge of their conditions or implementations,such as logic bombs that rely on uncertain conditions,let alone unsupported evasive techniques,which contain evasions without corresponding dismantling strategies and those leveraging unknown *** this paper,we present Antitoxin,a prototype for automatically exploring evasive *** utilizes multi-path exploration guided by taint analysis and probability calculations to effectively dismantle evasive *** probabilities of branch execution are derived from dynamic coverage,while taint analysis helps identify paths associated with evasive techniques that rely on uncertain ***,Antitoxin prioritizes branches with lower execution probabilities and those influenced by taint analysis for multi-path *** is achieved through forced execution,which forcefully sets the outcomes of branches on selected ***,Antitoxin employs active anti-evasion countermeasures to dismantle known evasive techniques,thereby reducing exploration ***,Antitoxin provides valuable insights into sensitive behaviors,facilitating deeper manual *** experiments on a set of highly evasive samples demonstrate that Antitoxin can effectively dismantle evasive techniques in a generic *** probability calculations guide the multi-path exploration of evasions without requiring prior knowledge of their conditions or implementations,enabling the dismantling of unsupported techniques such as C2 and signific

关键词： Malware analysis dynamic binary instrumentation forced execution taint analysis evasion detection

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：