检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Liu, Xiaoliang Shen, Fu-Rao Han, Feng Zhao, Jian Nie, Changhai State Key Laboratory for Novel Software Technology Nanjing University China Department of Computer Science and Technology Nanjing University China School of Artificial Intelligence Nanjing University China School of Electronic Science and Engineering Nanjing University China

Face recognition (FR) technology plays a crucial role in various applications, but its vulnerability to adversarial attacks poses significant security concerns. Existing research primarily focuses on transferability to different FR models, overlooking the direct transferability to victim's face images, which is a practical threat in real-world scenarios. In this study, we propose a novel adversarial attack method that considers both the transferability to the FR model and the victim's face image, called NerfTAP. Leveraging Neural Radiance Fields (NeRF)-based 3D Generative Adversarial Network (GAN), we generate new view face images for the source and target subjects to enhance transferability of adversarial patches. We introduce a style consistency loss to ensure the visual similarity between the adversarial UV map and the target UV map under a 0-1 mask, enhancing the effectiveness and naturalness of the generated adversarial face images. Extensive experiments and evaluations on various FR models demonstrate the superiority of our approach over existing attack techniques. Our work provides valuable insights for enhancing the robustness of FR systems in practical adversarial settings. © 2024, The Authors. All rights reserved.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Decentralized Counterfactual Value with Threat Detection in Multi-Agent Mixed Cooperative and Competitive Environment

SSRN

引用

SSRN 2023年

作者： Dong, Shaokang Li, Chao Yang, Shangdong Gao, Yang State Key Laboratory for Novel Software Technology Nanjing University China School of Computer Science Nanjing University of Posts and Telecommunications China

This paper introduces a fully decentralized method to address the problem of general mixed cooperation and competition in multi-agent setting. There are two main types of traditional Multi-Agent Reinforcement Learning (MARL) methods. The first kind is communication-based method which is designed to achieve coordination to deal with the non-stationary problem. However, it causes great communication overhead and can only be applied in fully cooperative environment. Another kind of methods follows Centralized Training and Decentralized Execution (CTDE) framework, which needs a centralized critic in the training process resulting in the huge computational complexity as the number of agents grows. In addition, neither approach is fully decentralized, because they either need to communicate the implicit information, or need to know the rewards of other agents. Different to the above methods, this paper introduces a Decentralized Counterfactual Value (DCV) to model other agents' behaviors, combined with the Threat Detection mechanism to learn the latent competitive or cooperative relationship with others. Finally, we prove that our method DCVTD does not affect the convergence of the basic algorithm and validate that DCVTD outperforms other fully decentralized methods, CTDE methods and other agent modeling or social influence methods in the experiments. © 2023, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Momentum Posterior Regularization for Multi-hop Dense Retrieval

arXiv

引用

arXiv 2024年

作者： Xia, Zehua Wu, Yuyang Xia, Yiyun Nguyen, Cam-Tu State Key Laboratory for Novel Software Technology Nanjing University China School of Computer Science Nanjing University China School of Artificial Intelligence Nanjing University China

Multi-hop question answering (QA) often requires sequential retrieval (multi-hop retrieval), where each hop retrieves missing knowledge based on information from previous hops. To facilitate more effective retrieval, we aim to distill knowledge from a posterior retrieval, which has access to posterior information like an answer, into a prior retrieval used during inference when such information is unavailable. Unfortunately, current methods for knowledge distillation in one-time retrieval are ineffective for multi-hop QA due to two issues: 1) Posterior information is often defined as the response (i.e. the answer), which may not clearly connect to the query without intermediate retrieval;and 2) The large knowledge gap between prior and posterior retrievals makes existing distillation methods unstable, even resulting in performance loss. As such, we propose MoPo (Momentum Posterior Regularization) with two key innovations: 1) Posterior information of one hop is defined as a query-focus summary from the golden knowledge of the previous and current hops;2) We develop an effective training strategy where the posterior retrieval is updated along with the prior retrieval via momentum moving average method, allowing smoother and effective distillation. Experiments on HotpotQA and StrategyQA demonstrate that MoPo outperforms existing baselines in both retrieval and downstream QA tasks‡ © 2024, CC BY-NC-SA.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

ATOM: Automated Black-Box Testing of Multi-Label Image Classification Systems

ATOM: Automated Black-Box Testing of Multi-Label Image Class...

引用

IEEE International Conference on Automated software Engineering (ASE)

作者： Shengyou Hu Huayao Wu Peng Wang Jing Chang Yongjun Tu Xiu Jiang Xintao Niu Changhai Nie Department of Computer Science and Technology State Key Laboratory for Novel Software Technology Nanjing University China Guangdong OPPO Mobile Telecommunications Corp. Ltd. China

Multi-label Image Classification Systems (MICSs) developed based on Deep Neural Networks (DNNs) are extensively used in people's daily life. Currently, although there are a variety of approaches to test DNN-based systems, they typically rely on the internals of DNNs to design test cases, and do not take the core specification of MICS (i.e., correctly recognizing multiple objects in a given image) into account. In this paper, we propose ATOM, an automated and systematic black-box testing framework for testing MICS. Specifically, ATOM exploits the label combination as the testing adequacy criteria, hoping to systematically examine the impact of correlations between a fixed number of labels on the classification ability of MICS. Then, ATOM leverages image search engine and natural language processing to find test images that are not only common to the real-world, but also relevant to target label combinations. Finally, ATOM combines metamorphic testing and label information to realize test oracle identification, based on which the ability of MICS in classifying different label combinations is evaluated. To evaluate the effectiveness of ATOM, we have performed experiments on two popular datasets of MICS, VOC and COCO (each with five state-of-the-art DNN models), and one real-world photo tagging application from our industrial partner. The experimental results reveal that the performance of current DNN-based MICSs remains less satisfactory even in recognizing correlations between only two labels, as ATOM triggers a total number of 6,049 such label combination related errors for all MICSs studied. In particular, ATOM reports 587 error-revealing images for the industrial MICS, in which 92% of them are confirmed by the developers.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Vulnerable Region-Aware Greybox Fuzzing

引用

Journal of computer science & technology 2021年第5期36卷 1212-1228页

作者： Ling-Yun Situ Zhi-Qiang Zuo Le Guan Lin-Zhang Wang Xuan-Dong Li Jin Shi Peng Liu State Key Laboratory for Novel Software Technology Nanjing UniversityNanjing 210023China School of Information Management Nanjing UniversityNanjing 210023China Department of Computer Science University of GeorgiaAthensGA 30602U.S.A. College of Information Sciences and Technology Pennsylvania State UniversityState CollegePA 16802U.S.A

Fuzzing is known to be one of the most effective techniques to uncover security vulnerabilities of large-scale software *** fuzzing,it is crucial to distribute the fuzzing resource appropriately so as to achieve the best fuzzing performance under a limited *** distribution strategies of American Fuzzy Lop(AFL)based greybox fuzzing focus on increasing coverage blindly without considering the metrics of code regions,thus lacking the insight regarding which region is more likely to be vulnerable and deserves more fuzzing *** tackle the above drawback by proposing a vulnerable region-aware greybox fuzzing ***,we distribute more fuzzing resources towards regions that are more likely to be vulnerable based on four kinds of code *** implemented the approach as an extension to AFL named ***-scale experimental evaluations validate the effectiveness and efficiency of RegionFuzz-11 new bugs including three new CVEs are successfully uncovered by RegionFuzz.

关键词： vulnerability detection greybox fuzzing code metrics resource distribution

来源：评论

学校读者我要写书评

暂无评论

Generalized Image Embedding for Multi-Domain Image Retrieval

Generalized Image Embedding for Multi-Domain Image Retrieval

引用

International Conference on computer Supported Cooperative Work in Design

作者： Boao Xiao Siyuan Wu Xin He Wanchun Dou State Key Laboratory for Novel Software Technology Nanjing University China School of Computer Science Nanjing University of Posts and Telecommunications China

Image embedding, being a fundamental task in computer vision, plays a crucial role in various downstream tasks such as image retrieval. Widely adopted in e-commerce and social media collaboration, image retrieval benefits greatly from representations learned by the embedding model. However, conventional embedding models are often trained on a single domain, leading to inadequate performance in the multi-domain scenario. To address this challenge, we introduce a generalized image embedding model designed for multi-domain image retrieval. The proposed method employs a contrastively learned Vision Transformer and a carefully crafted training scheme to enhance domain generalization capability. Our theoretical analysis and experimental results, conducted on a large-scale, real-world multi-domain image retrieval dataset, demonstrate the superiority of the proposed method over existing embedding models in terms of both accuracy and domain generalization capability.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Vehicle Check-In Data-driven POI Recommendation Based on Improved SVD and Graph Convolutional Network 19

Vehicle Check-In Data-driven POI Recommendation Based on Imp...

引用

2022 IEEE SmartWorld, 19th IEEE International Conference on Ubiquitous Intelligence and Computing, 2022 IEEE International Conference on Autonomous and Trusted Vehicles Conference, 22nd IEEE International Conference on Scalable Computing and Communications, 2022 IEEE International Conference on Digital Twin, 8th IEEE International Conference on Privacy Computing and 2022 IEEE International Conference on Metaverse, SmartWorld/UIC/ATC/ScalCom/DigitalTwin/PriComp/Metaverse 2022

作者： Liu, Yuwen Zhang, Jie Dou, Ruihan Zhou, Xiaokang Xu, Xiaolong Wang, Shoujin Qi, Lianyong College of Computer Science and Technology Qingdao China Nanjing University State Key Laboratory for Novel Software Technology Department of Computer Science and Technology China Shanghai Meteorological Information and Technology Support Center China University of Waterloo Faculty of Mathematics Waterloo Canada Shiga University Faculty of Data Science Japan Riken Center for Advanced Intelligence Project Riken Japan Nanjing University of Information Science and Technology School of Computer and Software China Macquarie University School of Computing Australia Qufu Normal University School of Computer Science China

ISBN: (纸本)9798350346558

The application of automobile tool software is becoming increasingly widespread, and its functions are becoming increasingly diverse. Using vehicle traffic data to give users with recommendations for Points-of-interest (POI) is becoming an intriguing research challenge. However, the current POI recommendation is either a sequence-based method, ignoring the interaction information between users and POIs. Or it needs a lot of model training time to use graph correlation method to mine the interaction between users or vehicles and POIs. In order to address these issues, we propose a method for vehicle check-in data-driven POI recommendation based on improved Singular Value Decomposition (SVD) and Graph Convolutional Network (GCN). First, we apply SVD to replace the neighborhood aggregation update method in traditional GCN, update the node embedding with fewer parameters and faster speed. Second, we propose an information enhancement strategy to make up for the information loss caused by the SVD. Experiments on large vehicle check-in datasets demonstrate that our method is both faster and more effective when making POI recommendations. © 2022 IEEE.

关键词： Singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

Camera Seismocardiogram Based Monitoring of Left Ventricular Ejection Time

引用

IEEE Transactions on Biomedical Engineering 2025年 PP卷 PP页

作者： Zhou, Zhiqin Huang, Jia Li, Haozhe Liu, Lin Zhu, Yingen Shan, Caifeng Wang, Wenjin Nanjing University State Key Laboratory for Novel Software Technology School of Intelligence Science and Technology China Southern University of Science and Technology Department of Biomedical Engineering China The Third People's Hospital of Shenzhen Intensive Care Unit China Shandong University of Science and Technology College of Electrical Engineering and Automation China

Left Ventricular Ejection Time (LVET), reflecting the duration from the onset to the end of blood ejection by the left ventricle during each heartbeat, is a critical parameter for measuring cardiac pumping efficiency. Continuous and regular monitoring of LVET is particularly crucial in assessing cardiac health, valvular function, and myocardial contractility. Seismocardiogram (SCG) signals can be utilized for LVET monitoring, as the temporal distance between the aortic valve opening (AO) and aortic valve closure (AC) in SCG signals can accurately depict LVET. This study proposes a novel way to extract LVET from laser speckle videos recorded by a remote camera based on the principle of defocused speckle imaging, thereby enabling non-contact monitoring of LVET. We extract both the low-frequency components of laser speckle motion (LSM-LF), regarded as SCG signals, and the high-frequency components of laser speckle motion (LSM-HF) from recorded videos. We utilize LSM-HF to assist the detection of AO and AC markers in LSM-LF. We validated the effectiveness of our AO and AC detection algorithm on a self-made dataset comprising 21 participants with 9616 SCG cycles. The benchmark shows that the detection accuracy for AO and AC reached 98.16% and 97.94%, respectively, with an mean absolute error of 0.5571 ms for LVET estimation. The results demonstrate that camera-SCG has strong potential for cardiac health monitoring. © 1964-2012 IEEE.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Duplicate Data Elimination Optimization Strategy Based on Edge Service Management

Duplicate Data Elimination Optimization Strategy Based on Ed...

引用

Cyber science and technology Congress (CyberSciTech), IEEE

作者： Yonglin Pu Xiaolong Xu Lianyong Qi Wanchun Dou Jiong Yu Ziyang Li School of Software Nanjing University of Information Science and Technology Nanjing China College of Computer Science and Technology China University of Petroleum (East China) Dongying China State Key Laboratory for Novel Software Technology Nanjing University Nanjing China School of Software Xinjiang University Urumqi China

ISBN: (数字)9798331532093

ISBN: (纸本)9798331532109

Due to the absence of duplicate data elimination technologies in edge architectures, the interaction experience with mobile users suffers. In response to this issue, an optimized strategy for duplicate data elimination, based on edge service management (DDE-ESM), is proposed. Initially, a model for the placement of edge devices and a model for the collection of cache data are established. Subsequently, a caching device placement algorithm is designed; this not only undertakes comprehensive data collection and cache storage but also forms the basis for querying duplicate data within the architecture. Following this, a model for querying duplicate data in the cache is created, identifying duplicates by segmenting data blocks and generating data fingerprints via hash functions. Ultimately, a model for eliminating duplicate data at the edge is developed, which leads to the creation of an algorithm that rationally deletes duplicate data in accordance with device resource constraints. Experimental results indicate that this optimization strategy for duplicate data elimination not only saves resources on edge devices but also significantly enhances data management efficiency and the utilization of architectural resources.

关键词： Hash functions Fingerprint recognition Data collection Data models Cache storage Optimization

来源：评论

学校读者我要写书评

暂无评论

CACam: Consecutive Angular Measurements with Camera on Smartphone for Distance Estimation

CACam: Consecutive Angular Measurements with Camera on Smart...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Yifan Zhu Xiaojun Zhu College of Computer Science and Technology Nanjing University of Aeronautics and Astronautics Nanjing China State Key Laboratory of Novel Software Technology Nanjing University Nanjing China

To estimate the distance between a user and the target object for personal interest or for subsequent actions, the commonly used method is to search the target in the map application. However this is limited by GPS signal status under different environments and incomplete information on the map application of smartphones. This paper introduces a ranging method, CACam, which utilizes consecutive angle measurement based on camera of smartphone. Geometric model and ranging process are proposed to calculate the distance. When the user aligns the target from different position points, we record readings from the gyroscope sensor during every rotation in the procedure. The camera is used for aligning the target to build geometric model. Vector mapping of angular readings and data processing of intermediate points are used to get more accurate angles, which is suitable for the situation that the smartphone could not be strictly rotated in perpendicular to the plane of the gravity. The effectiveness of our ranging method is verified through experiments in outdoor environments.

关键词： Geometric modeling Cameras Search problems Distance measurement Hardware Gyroscopes Smart phones

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：