检索结果-内蒙古大学图书馆

IEEE International Conference on Multimedia and Expo (ICME)

作者： Ying Ren Kailai Shen Zhe Ye Diqun Yan Faculty of Electrical Engineering and Computer Science Ningbo University Ningbo China Juphoon System Software Co. Ltd Ningbo China Key Laboratory of Computing Power Network and Information Security Ministry of Education Qilu University of Technology (Shandong Academy of Sciences) China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Non-Intrusive speech quality assessment (NISQA) has gained significant attention for predicting speech’s mean opinion score (MOS) without requiring the reference speech. Researchers have gradually started to apply NISQA to various practical scenarios. However, little attention has been paid to the security of NISQA models. Backdoor attacks represent the most serious threat to deep neural networks (DNNs) due to the fact that backdoors possess a very high attack success rate once embedded. However, existing backdoor attacks assume that the attacker actively feeds samples containing triggers into the model during the inference phase. This is not adapted to the specific scenario of NISQA. And current backdoor attacks on regression tasks lack an objective metric to measure the attack performance. To address these issues, we propose a novel backdoor triggering approach (EventTrojan) that utilizes an event during the usage of the NISQA model as a trigger. Moreover, we innovatively provide an objective metric for backdoor attacks on regression tasks. Extensive experiments on four benchmark datasets demonstrate the effectiveness of the EventTrojan attack. Besides, it also has good resistance to several defense methods.

关键词： Measurement Resistance Adaptation models Multimedia systems Current measurement Artificial neural networks Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

SML: A Backdoor Defense for Non-Intrusive Speech Quality Assessment via Semi-Supervised and Multi-Task Learning

SML: A Backdoor Defense for Non-Intrusive Speech Quality Ass...

引用

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

作者： Ying Ren Wenjie Zhang Jiahong Ye Jie Li Diqun Yan Bin Ma Faculty of Electrical Engineering and Computer Science Ningbo University Ningbo China College of Digital Technology and Engineering Ningbo University of Finance and Economics Ningbo China Key Laboratory of Computing Power Network and Information Security Ministry of Education Qilu University of Technology (Shandong Academy of Sciences) China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Non-intrusive speech quality assessment (NISQA) is widely used in speech downstream tasks due to its ability to predict the quality of speech without a reference speech. However, few researchers have focused on the backdoor security of NISQA. Despite the backdoor defenses have been extensively studied to mitigate the threat of maliciously modifications in deep neural networks. In particular, semi-supervised based backdoor defenses have excellent defensive performance by depriving backdoor attacks of their most essential need. But these defense methods rely on data-augmentation consistency and thus cannot be applied to NISQA. In this work, we propose a backdoor defense based on semi-supervised and multi-task learning (SML). Semi-supervised learning is based on the simple assumption that the same input should be as consistent as possible in two similar models. Multi-task learning further improves the prediction performance of mean opinion score (MOS) by learning the tasks of perceptual evaluation of speech quality (PESQ), short-time objective intelligibility (STOI) and speech distortion index (SDI). Extensive experiments involving five backdoor defenses against five backdoor attacks on two benchmark datasets demonstrate the superiority of our SML approach.

关键词： Acoustic distortion Artificial neural networks Semisupervised learning Benchmark testing Multitasking Acoustics Quality assessment security Indexes Speech processing

来源：评论

学校读者我要写书评

暂无评论

Unleashing the Potential of Decentralized Federated Learning in Healthcare: A Comprehensive Survey on Insights and Challenges

Unleashing the Potential of Decentralized Federated Learning...

引用

IEEE International Conference on Data Science in Cyberspace (DSC)

作者： Haibo Cheng Wenjian Liu Youyang Qu Kexue Fu Qinyuan Wang Faculty of Data Science City University of Macau Macau China Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center (National Supercomputer Center in Jinan) Qilu University of Technology (Shandong Academy of Sciences) Jinan China

ISBN: (数字)9798350391367

ISBN: (纸本)9798350391374

Federated learning (FL) is a distributed learning method that reduces data transmission and privacy risks. However, the existence of a central server may become a target of attackers, resulting in an increased risk of user privacy leakage. The emergence of Decentralized Federated Learning (DFL) is posited to enhance data privacy and system reliability by reducing reliance on central servers and enhancing local data control, especially in the medical field, where personal health data is protected. This research investigates the application of (DFL) in the medical domain, focusing on its limitations. This study employs a systematic literature review, adhering to PRISMA guidelines, to comprehensively understand DFL’s current landscape and challenges in healthcare. The key findings highlight four principal obstacles faced by DFL: security and privacy, communication efficiency, data and model heterogeneity, and incentive mechanisms. It also explores the potential for DFL to facilitate personalized healthcare treatments by large datasets distributed across various medical institutions. This research fills a significant void by systematically reviewing DFL technologies in healthcare. It provides insights into applications, challenges, and future directions for participants seeking secure and efficient healthcare data handling.

关键词： Surveys Data privacy Systematics Federated learning Medical services Data models Servers security Reliability Systematic literature review

来源：评论

学校读者我要写书评

暂无评论

Speed Master: Quick or Slow Play to Attack Speaker Recognition 39

Speed Master: Quick or Slow Play to Attack Speaker Recogniti...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Ye, Zhe Zhang, Wenjie Ren, Ying Kang, Xiangui Yan, Diqun Ma, Bin Wang, Shiqi Guangdong Key Lab of Information Security School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Faculty of Electrical Engineering and Computer Science Ningbo University Ningbo China Key Laboratory of Computing Power Network and Information Security Ministry of Education Qilu University of Technology Shandong Academy of Sciences Jinan China Department of Computer Science City University of Hong Kong Hong Kong

ISBN: (纸本)157735897X

Backdoor attacks pose a significant threat during the model’s training phase. Attackers craft pre-defined triggers to break deep neural networks, ensuring the model accurately classifies clean samples during inference yet erroneously classifies samples added with these triggers. Recent studies have shown that speaker recognition systems trained on large-scale data are susceptible to backdoor attacks. Existing attackers employ unnoticed ambient sounds as triggers. However, these sounds are not inherently part of the training samples themselves. In essence, triggers can be designed to maintain an intrinsic connection with the original speech to enhance stealthiness. Our paper presents a novel attack methodology named Speed Master, which undermines deep neural networks by manipulating the speed of speech samples. Specifically, we execute poison-only backdoor attacks using speed or tempo adjustment. Changes in speech rate have become a common occurrence, as seen on platforms that allow users to adjust playback speed. In real-world scenarios, people naturally adjust their speaking rate depending on the context. As a result, changes in a speaker’s speech rate are typically perceived as normal and are unlikely to raise suspicion. Furthermore, detecting such subtle adjustments becomes challenging for users without reference speech. Our comprehensive experiments demonstrate that Speed Master can achieve an ASR over 99% in the digital domain, with only a 0.6% poisoning rate. Additionally, we validate the feasibility of Speed Master in the real world and its resistance to typical defensive measures. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

SRIF: Data-Free Knowledge Distillation via Stable Regulation and Input Filtering

引用

IEEE Transactions on Multimedia 2025年

作者： Zhang, Zherui Xu, Rongtao Wang, Changwei Xu, Shibiao Zhou, Jie Xu, Wenhao Gao, Longxiang Xu, Wenbo Guo, Li Beijing University of Posts and Telecommunications School of Artificial Intelligence Beijing China Chinese Academy of Sciences State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Beijing China Qilu University of Technology Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computing Power Internet and Service Computing Jinan China

Data-free knowledge distillation (DFKD) enables knowledge transfer from a pre-trained teacher to a student network without accessing the real dataset. However, generator-based DFKD methods struggle to ensure that the synthetic images accurately reflect the real dataset distribution. The update of the generator network relies heavily on teacher category guidance, but varying teacher prediction accuracy across categories leads to inconsistent synthetic image quality. Such variations introduce a distribution shift between synthetic and real datasets, negatively impacting student network performance during knowledge distillation. To address this challenge, we propose the SRIF, comprising two components: Student-Driven Flexible Filtering (SDFF) and Re-weighting for Independent Regularization (RIR). SDFF filters out synthetic images affected by the category distribution shift during data generation, producing a more reliable dataset. RIR, applied during distillation, encourages the student to learn stable causal relationships through sample reweighting. Both components flexibly integrate into existing DFKD frameworks, improving performance while reducing training costs. © 1999-2012 IEEE.

关键词： Photointerpretation

来源：评论

学校读者我要写书评

暂无评论

Improved YOLOv5 Object Detection Method for Cervical Fluid Cell Pathological Image Detection

Improved YOLOv5 Object Detection Method for Cervical Fluid C...

引用

International Joint Conference on Neural networks (IJCNN)

作者： Qingyan Ding Xuecheng Dong Na Li Yu Pan Wan Zheng Dept. Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center (National Supercomputer Center in Jinan) Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Cervical liquid-based cell pathology image detection and recognition is a key task in gynecological pathology aimed at early detection of cervical precancerous lesions. In this study, a deep learning based method is proposed for the detection and recognition of cervical liquid-based cell images. This paper is based on the YOLOv5 network model for improvement. Firstly, in the data preprocessing stage: for the characteristics of the dataset, this paper proposes two data enhancement methods to augment the training dataset; secondly, the Transformer module for local information enhancement is improved by introducing a local attention mechanism, which restricts each position to only pay attention to positions within a certain range around it; then, the spatial pyramid pooling module (DSPPCSPC) is improved, and the spatial Pyramid Pooling Module (DSPPC-SPC), introducing deformable mask of deformable convolution (ODConv); then, this paper proposes a new Attention Module (XCAttention), improving S2Attention to focus on the local information of the image, and introducing its Channel Attention Module into the third branch, which performs the Dot with the first and the second branches that have gone through Spatial-shift to obtain the final output. Product operation to get the final output. Through experiments, this paper’s algorithm improves in Precision, Recall, mAP50, and mAP50-95 metrics by 1.9%, 1.9%, 2%, and 1.2%, respectively. The experimental results show that this paper’s algorithm can effectively improve the detection performance of the model.

关键词： YOLO Training Deep learning Pathology Technological innovation Image recognition Convolution

来源：评论

学校读者我要写书评

暂无评论

Automatic recognition of depression based on audio and video:A review

引用

World Journal of Psychiatry 2024年第2期14卷 225-233页

作者： Meng-Meng Han Xing-Yun Li Xin-Yu Yi Yun-Shao Zheng Wei-Li Xia Ya-Fei Liu Qing-Xiang Wang Shandong Mental Health Center Shandong UniversityJinan 250014Shandong ProvinceChina Key Laboratory of Computing Power Network and Information Security Ministry of EducationShandong Computer Science Center(National Supercomputer Center in Jinan)Qilu University of Technology(Shandong Academy of Sciences)Jinan 250353Shandong ProvinceChina Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and TechnologyQilu University of Technology(Shandong Academy of Sciences)Jinan 250353Shandong ProvinceChina Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer ScienceJinan 250353Shandong ProvinceChina Department of Ward Two Shandong Mental Health CenterShandong UniversityJinan 250014Shandong ProvinceChina

Depression is a common mental health *** current depression detection methods,specialized physicians often engage in conversations and physiological examinations based on standardized scales as auxiliary measures for depression ***-biological markers-typically classified as verbal or non-verbal and deemed crucial evaluation criteria for depression-have not been effectively *** physicians usually require extensive training and experience to capture changes in these *** in deep learning technology have provided technical support for capturing non-biological *** researchers have proposed automatic depression estimation(ADE)systems based on sounds and videos to assist physicians in capturing these features and conducting depression *** article summarizes commonly used public datasets and recent research on audio-and video-based ADE based on three perspectives:Datasets,deficiencies in existing research,and future development directions.

关键词： Depression recognition Deep learning Automatic depression estimation System Audio processing Image processing Feature fusion Future development

来源：评论

学校读者我要写书评

暂无评论

Flexible and Secure Access Control for Ehr Sharing Based on Blockchain

SSRN

引用

SSRN 2023年

作者： Ma, Haobin Zhou, Dehua Li, Peng College of Information Science and Technology Jinan University Guangzhou510632 China Guangdong Provincial Key Laboratory of Power System Network Security Guangzhou510663 China

With the rapid development of the healthcare industry, many Electronic Health Records (EHRs) have emerged in different healthcare centers. However, lots of personal medical data is stored directly in plaintext at a single healthcare center, which makes it suffer from the single point of failure and bring many security and privacy issues such as privacy information leakage or tampering. The combination of blockchain and attributed-based cryptography can effectively solve the above problems. Therefore, in order to enable flexible fine-grained access control and efficient data retrieval while achieving privacy preserving, we employ the ciphertext-policy attributed-based encryption (CP-ABE) and ciphertext-policy attribute-based searchable encryption (CP-ABSE) to propose a hidden policy attribute-based access control scheme. In addition, we combine consortium blockchain and smart contract to provide secure and reliable search and outsourcing decryption. Finally, through the security analysis and experimental results, we demonstrate that our scheme is secure and efficient. © 2023, The Authors. All rights reserved.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Traffic Anomaly Detection Model of Electric power Industrial Control Based on DBN-LSTM 23

Traffic Anomaly Detection Model of Electric Power Industrial...

引用

23rd IEEE International Conference on High Performance computing and Communications, 7th IEEE International Conference on Data Science and Systems, 19th IEEE International Conference on Smart City and 7th IEEE International Conference on Dependability in Sensor, Cloud and Big Data Systems and Applications, HPCC-DSS-SmartCity-DependSys 2021

作者： Zhang, Shijun Lai, Ji Yao, Qigui Platform Operation and Safety Center Information and Communication Branch of State Grid Jibei Electric Power Co. Ltd Beijing China Global Energy Interconnection Research Institute Co. Ltd State Grid Key Laboratory of Information & Network Security Nanjing China

ISBN: (纸本)9781665494571

With the construction of smart city and smart grid, energy Internet is booming, and the deployment of smart power terminals has increased, with a wide range of deployment locations, and the level of intelligence of the terminals has increased. However, most of the intelligent terminals are deployed in the non-controlled industrial control network environment, which brings new challenges to the network security. Therefore, the traffic anomaly detection of electric power industrial control becomes the first step to ensure the safety of the peripheral network. Currently, the traffic anomaly detection system of electric power industrial control based on machine learning has the problems of low computing efficiency and weak representation ability. To solve the problem of low efficiency in traditional deep learning, the deep belief network (DBN) is used to extract the non-linear features of data, which can improve the time efficiency to a certain extent. To solve the problem that the representation ability of traditional machine learning is weak, long short term memory network (LSTM) is used to improve the representation ability of traditional shallow machine learning. The experimental results suggest that the method is better than the traditional machine learning method in accuracy, and the recall rate reaches up to 97.3%, AUC runs up to 0.927. © 2021 IEEE.

关键词： Smart cities Industrial control network security Feature extraction power systems Smart grids Safety

来源：评论

学校读者我要写书评

暂无评论

Metadata modeling and management for power communication network operation and maintenance 10

Metadata modeling and management for power communication net...

引用

10th IEEE Joint International information Technology and Artificial Intelligence Conference, ITAIC 2022

作者： Zhu, Lipeng Xu, Menghan Zhou, Aihua Qiao, Junfeng State Grid Smart Grid Research Institute Co. Ltd Beijing China State Grid Jiangsu Electric Power Co. LTD. Information & Telecommunication Branch Nanjing China State Grid Key Laboratory of Information & Network Security Nanjing China

ISBN: (数字)9781665422079

ISBN: (纸本)9781665422079

With the rapid development of the smart grid, the network scale and structure of the power grid are becoming larger and more complex, and the corresponding power communication network operation and maintenance management has put forward higher requirements. At present, the power communication network operation and maintenance data face the following two major problems: the metadata model is not unified, and there is a lack of management tools for the operation and maintenance metadata. Aiming at these problems, this paper establishes a unified power communication network operation and maintenance meta-model, and designs a power communication network operation and maintenance metadata management system based on this, which realizes the automatic acquisition of metadata and the analysis of blood relationship information. The work of this paper not only provides a unified information model specification for the management of the metadata of the power communication network operation and maintenance, which facilitates the storage and interaction of the metadata, but also provides a useful technical reference for the management of the metadata of the power communication network operation and maintenance. © 2022 IEEE.

关键词： Metadata

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：