检索结果-内蒙古大学图书馆

12th International Conference on Advanced Cloud and Big Data, CBD 2024

作者： Ding, Shiru Wang, Zhen Lu, Yifei School of Electrical and Automation Engineering Nanjing Normal University China School of Computer Science and Engineering Nanjing University of Science and Technology China

ISBN: (纸本)9798331511074

Previous deep learning-based Network Intrusion Detection Systems (NIDS) require a sufficient number of labeled samples to train deep neural network models. However, in certain scenarios of the Internet of Things (IoT), such as zero-day attacks, abnormal data is scarce and cannot meet the training conditions for neural network models. Thus, there is a need for a NIDS capable of few-shot learning. Additionally, previous online network intrusion detection systems did not comprehensively consider the significant impact of data feature drift and catastrophic forgetting on the performance of online models. In order to solve these two problems, we designed an adaptive online few-shot network intrusion detection system based on meta-learning. We evaluated the proposed system using the public datasets CIC-IDS2017, and the results show that our system can effectively handle the challenge of small samples while continuously adapting to new data with minimal catastrophic forgetting. © 2024 IEEE.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Reversible Data Hiding in Shared Images Based on Syndrome Decoding and Homomorphism

引用

IEEE Transactions on Cloud Computing 2023年第3期11卷 3085-3098页

作者： Xiong, Lizhi Han, Xiao Yang, Ching-Nung Zhang, Xinpeng Nanjing University of Information Science and Technology School of Computer Science Jiangsu Nanjing210044 China Nanjing University of Information Science and Technology Engineering Research Center of Digital Forensics Ministry of Education CICAEET Jiangsu Nanjing210044 China National Dong Hwa University Department of Computer Science & Information Engineering Hualien974 Taiwan Fudan University School of Computer Science Shanghai200433 China

Reversible Data Hiding in Encrypted Images (RDHEI) has drawn increasing concern in multimedia cloud computing scenarios. It embeds secret message into the encrypted carrier while preserving the confidentiality of the image. However, most RDHEI schemes have only one hider and one image carrier which are not efficient for the distributed system with multiple participants. Moreover, if the stego-image is lost, the cover image and the embedded data cannot be restored. To solve above issues, this paper proposes Reversible Data Hiding in Shared Images based on syndrome decoding and homomorphism (RDHSI). In RDHSI, the cover image and secret data are distributed using secret sharing. Multiple data hiders embed shared data into the shared images based on the Syndrome Decoding of Hamming codes and the additive homomorphism of the polynomial. On the receiver side, the lossless cover image and secret data are obtained. The reversibility and privacy preserving of the cover image, the complete extraction of secret data, and the fault-tolerance of the proposed schemes are achieved. The experimental results show that PSNR of the marked image is above 54 dB at the embedding rate of 0.43 bpp. In order to improve the payload of secret data, the Modified RDHSI is further proposed. Generally, the proposed schemes are secure, effective and fault-tolerant. © 2013 IEEE.

关键词： Data privacy

来源：评论

学校读者我要写书评

暂无评论

QFSM: A Novel Quantum Federated Learning Algorithm for Speech Emotion Recognition With Minimal Gated Unit in 5G IoV

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年第10期9卷 1-12页

作者： Qu, Zhiguo Chen, Zhixiao Dehdashti, Shahram Tiwari, Prayag Jiangsu Collaborative Innovation Center of Atmospheric Environment the Equipment Technology and the Engineering Research Center of Digital Forensics Ministry of Education and the School of Computer Science Nanjing University of Information Science and Technology Nanjing China School of Computer Science Nanjing University of Information Science and Technology Nanjing China Emmy-Noether Gruppe Theoretisches Quantensystemdesign Lehrstuhl Für Theoretische Informationstechnik Technische Universität München Munich Germany School of Information Technology Halmstad University Halmstad Sweden

The technology of speech emotion recognition (SER) has been widely applied in the field of human-computer interaction within the Internet of Vehicles (IoV). The incorporation of emerging technologies such as artificial intelligence and big data has accelerated the advancement of SER technology. However, this reveals challenges such as limited computational resources, data processing inefficiency, and security and privacy concerns. In recent years, quantum machine learning has been applied to the field of intelligent transportation, which has demonstrated its various advantages, including high prediction accuracy, robust noise resistance, and strong security. This study first integrates quantum federated learning (QFL) into 5G IoV using a quantum minimal gated unit (QMGU) recurrent neural network for local training. Then, it proposes a novel quantum federated learning algorithm, QFSM, to further enhance computational efficiency and privacy protection. Experimental results demonstrate that compared to existing algorithms using quantum long short-term memory network or quantum gated recurrent unit models, the QFSM algorithm has a higher recognition accuracy and faster training convergence rate. It also performs better in terms of privacy protection and noise robustness, enhancing its applicability and practicality. IEEE

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

MPTrack: Visual Tracking with Reliable Motion Prompts 24

MPTrack: Visual Tracking with Reliable Motion Prompts

引用

2024 International Conference on Artificial Intelligence of Things and Computing, AITC 2024

作者： Sun, Songtao Li, Lunbo Aerospace System Engineering of Shanghai Shanghai China School of Computer Science and Engineering Nanjing University of Science and Technology Jiangsu Nanjing China

ISBN: (纸本)9798400709869

Recently, the attention mechanism has been introduced into object tracking, making significant improvements in tracking performance. However, the tracking target often undergoes deformation during tracking, which can easily lead to the loss of the target. In this paper, we design the motion perception unit (MPU) to enhance tracking stability. The MPU generates motion prompts rich in dynamic target appearance information and integrates them into object modeling. This integration allows our tracker to better adapt to changes in target motion and appearance, improving the adaptability to the target motion. Extensive experiments demonstrate that our tracker MPTrack outperforms all state-of-the-art trackers on five widely recognized benchmarks, including GOT-10k, TrackingNet, LaSOText, UAV123, and TNL2K, while running at real-time speed. Copyright © 2024 held by the owner/author(s). Publication rights licensed to ACM.

关键词： Target motion

来源：评论

学校读者我要写书评

暂无评论

Generating Image Adversarial Example by Modifying JPEG Stream 24

Generating Image Adversarial Example by Modifying JPEG Strea...

引用

2024 International Conference on computer Vision and Deep Learning, CVDL 2024

作者： Zhao, Junjie Qiao, Sen Wang, Jinwei Liu, Guangjie School of Electronics and Information Engineering Nanjing University of Information Science and Technology Jiangsu Nanjing China School of Computer Science Nanjing University of Information Science and Technology Jiangsu Nanjing China

ISBN: (纸本)9798400718199

Research on adversarial attacks mainly focuses on reducing the amplitude of disturbances, increasing the success rate of attacks, and improving attack efficiency. However, the adversarial examples are all in the form of raw (float matrix or lossless compressed pictures). Due to the difficulty of storage and transmission caused by a large amount of redundant information, raw pictures rarely exist in the real world. In this work, a convolution operation is used to simulate the JPEG decompression process, forming a decompression module independent of the classification model to obtain the gradient information about the JPEG stream. We establish a process for generating JPEG adversarial examples. To improve the decoded image's visual quality and reduce the perturbation amplitude, we select the complex texture area according to the human visual model and select only the most effective points in each DCT block to embed adversarial noise. A large number of tests on the ImageNet dataset show the effectiveness of the proposed method. © 2024 ACM.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

M3Rec: Selective State Space Models with Mixture-of-Modality Experts for Multi-Modal Sequential Recommendation

M3Rec: Selective State Space Models with Mixture-of-Modality...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Guo, Xu Zhang, Tong Xue, Yufei Wang, Chenxu Wang, Fuyun Cui, Zhen School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China School of Mechanical Engineering Shanghai Jiao Tong University Shanghai China

ISBN: (纸本)9798350368741

The rapid growth of multimedia-sharing platforms drives the development of recommender systems. While traditional ID-based methods for mining user behavior signals are well-studied, research into multimodal sequential recommendation remains nascent. Current approaches face three critical challenges: (1) inadequate modeling of user preferences across diverse modalities, (2) ineffective capture of user action sequence dependencies hinders representation learning of preferences, and (3) inefficiency in Transformer-based models due to the quadratic complexity of attention mechanisms. To address these issues, we propose M3Rec, a Mamba-based selective state space model incorporating Mixture-of-Modality experts for Multimodal sequential recommendation. M3Rec strengthens the modeling of user action sequence dependencies through shared Mamba blocks across modalities and employs modality experts to extract modality-specific user preferences. The shared Mamba blocks efficiently model long-term user preferences with fast inference and linear scalability through hardware-aware parallelism, enhancing ID-based sequence signals and filtering out non-action-dependent redundant information. This enables more accurate modeling of user preferences across heterogeneous data. Extensive experiments on three public datasets validate the model's effectiveness. The implementation is released at https://***/Xu107/M3Rec-main. © 2025 IEEE.

关键词： Mamba Multimedia Sequential Recommendation

来源：评论

学校读者我要写书评

暂无评论

Intelligent media computing technology and application for media convergence

引用

CAAI Transactions on Intelligence technology 2022年第3期7卷 329-330页

作者： Zechao Li School of Computer Science and Engineering Nanjing University of Science and TechnologyNanjingChina

Media Convergence is the merging of mass communication outlets—print,television,radio,and the Internet—along with portable and interactive technologies through various digital media *** a new influential mainstream media,Media Convergence has now become a national strategy to integrate multiple media forms into one ***,the intelligent media computing technology and application,including 5G,Augmented Reality/Visual Reality,Natural Language Processing,computer Vision,Robotics,Big data,and Machine/Deep/Reinforcement/Transfer learning,should evolve into a knowledge base for purposes of delivering a dynamic experience and innovating media communication ***,how does one integrated media provide effective algorithm structures and tools that could merge,transform,and process various media forms,that is,crossmodal/multi-modal learning and representation?This question remains to be answered.

关键词： computing application media

来源：评论

学校读者我要写书评

暂无评论

Provable Data Possession with Outsourced Tag Generation for AI-Driven E-Commerce

引用

computers, Materials & Continua 2025年第5期83卷 2719-2734页

作者： Yi Li Wenying Zheng Yu-Sheng Su Meiqin Tang School of Computer Science Nanjing University of Information Science and TechnologyNanjing210044China School of Computer Science and Technology(School of Artificial Intelligence) Zhejiang Sci-Tech UniversityHangzhou310018China Department of Computer Science and Information Engineering National Chung Cheng UniversityChiayi621301TaiwanChina Advanced Institute of Manufacturing with High-tech Innovations National Chung Cheng UniversityChiayi621301TaiwanChina Department of Computer Science and Engineering Taiwan Ocean UniversityKeelung202301TaiwanChina School of Integrated Circuits Wuxi Vocational College of Science and TechnologyWuxi214028China

AI applications have become ubiquitous,bringing significant convenience to various *** e-commerce,AI can enhance product recommendations for individuals and provide businesses with more accurate predictions for market strategy ***,if the data used for AI applications is damaged or lost,it will inevitably affect the effectiveness of these AI ***,it is essential to verify the integrity of e-commerce *** existing Provable Data Possession(PDP)protocols can verify the integrity of cloud data,they are not suitable for e-commerce scenarios due to the limited computational capabilities of edge servers,which cannot handle the high computational overhead of generating homomorphic verification tags in *** address this issue,we propose PDP with Outsourced Tag Generation for AI-driven e-commerce,which outsources the computation of homomorphic verification tags to cloud servers while introducing a lightweight verification method to ensure that the tags match the uploaded ***,the proposed scheme supports dynamic operations such as adding,deleting,and modifying data,enhancing its ***,experiments show that the additional computational overhead introduced by outsourcing homomorphic verification tags is acceptable compared to the original PDP.

关键词： Provable data possession data auditing cloud computing e-commerce bloom filter

来源：评论

学校读者我要写书评

暂无评论

Autonomous Exploration Algorithm for Mobile Robots in Unknown Confined Environment 16

Autonomous Exploration Algorithm for Mobile Robots in Unknow...

引用

16th International Conference on Intelligent Human-Machine Systems and Cybernetics, IHMSC 2024

作者： Lyu, Zhixiao Yin, Yue Liu, Qinyang Yang, Tianyao Nanjing Normal University School of Electrical and Automation Engineering China Nanjing University of Information Science & Technology School of Computer Science School of Cyber Science and Engineering China

ISBN: (纸本)9798331540241

For the efficiency and precision of autonomous exploration, this paper proposes an optimized algorithm for mobile robots in unknown confined environment. However, the traditional algorithms frequently lead to excessive navigation distance and waste of exploration time consumption. To address these limitations, frontier point selection is improved by utilizing RRT, accomplishing significant improvement in computational efficiency. The exploration path planning was accomplished through the utilization of an enhanced Dijkstra's algorithm, in conjunction with a topological map. Through experimental comparisons, the proposed algorithm achieves significant advantages over existing frontier-based and RRT-based algorithms when dealing with complex environments. In particular, our system is able to reduce the traveling distance by up to 55.16% in difficult scenarios. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Meta-learning with orthogonal softmax layer (MLOSL) for small sample hyperspectral image classification

引用

Multimedia Tools and Applications 2025年 1-30页

作者： Amoako, Prince Yaw Owusu Cao, Guo Shi, Hao Arthur, John Kingsley Agyenim, Yaw Oti Boateng School of Computer Science and Engineering Nanjing University of Science and Technology Xiaolingwei Nanjing City210094 China Department of Computer Science Valley View University Accra Oyibi Ghana ETSU Municipal Center East Tennessee State University Johnson CityTN United States

Recent research on Hyperspectral image classification (HSIC) has been much concerned with small sample size due to the challenges of obtaining large labeled data. Coupled with this is the challenge with optimization of the models’ structure and computational complexities. Meta-learning techniques have been proposed to improve classification accuracy on small training data samples. The existing meta-learning techniques still require several pre-collected labeled hyperspectral images (HSI) source datasets of large labeled samples for meta-training, which is undoubtedly a huge task and time-consuming. We propose a meta-learning with orthogonal softmax layer (MLOSL) for small sample HSIC. Orthogonality is introduced to address the structural and computational complexities of deep convolutional neural network (CNN) models for HSIC. An unsupervised meta-learning task construction extracts multiview spectral dimension features of different bands from the same sample, whereas the spatial dimension multiview features are extracted through data augmentation. The orthogonal softmax layer captures a wider variety of patterns in the data by promoting variation among the weight vectors, potentially improving generalization to unobserved samples and minimizing computational time. Label smoothing is added to the classification loss to lessen the impact of incorrectly classified samples and class imbalance. Experimental outcomes on four target datasets compared to existing state-of-the-art models indicate the competitiveness of the proposed model's small sample HSIC performance and its suitability in time-constraint applications. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

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：