检索结果-内蒙古大学图书馆

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

作者： Yingjie Ma Zitong Yu Xun Lin Weicheng Xie Linlin Shen College of Computer Science and Software Engineering Shenzhen University Great Bay University National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Guangdong Provincial Key Laboratory of Intelligent Information Processing

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In the domain of facial recognition security, multimodal Face Anti-Spoofing (FAS) is essential for countering presentation attacks. However, existing technologies encounter challenges due to modality biases and imbalances, as well as domain shifts. Our research introduces a Mixture of Experts (MoE) model to address these issues effectively. We identified three limitations in traditional MoE approaches to multimodal FAS: (1) Coarse-grained experts’ inability to capture nuanced spoofing indicators; (2) Gated networks’ susceptibility to input noise affecting decision-making; (3) MoE’s sensitivity to prompt tokens leading to overfitting with conventional learning methods. To mitigate these, we propose the Bypass Isolated Gating MoE (BIG-MoE) framework, featuring: (1) Fine-grained experts for enhanced detection of subtle spoofing cues; (2) An isolation gating mechanism to counteract input noise; (3) A novel differential convolutional prompt bypass enriching the gating network with critical local features, thereby improving perceptual capabilities. Extensive experiments on four benchmark datasets demonstrate significant generalization performance improvement in multimodal FAS task. The code is released at https://***/murInJ/BIG-MoE.

关键词： Convolutional codes Sensitivity Convolution Face recognition Noise Logic gates Feature extraction Security Speech processing Overfitting

来源：评论

学校读者我要写书评

暂无评论

HiRes: Hierarchical Feature Optimization and Rescorer for Automatic ICD Coding

HiRes: Hierarchical Feature Optimization and Rescorer for Au...

引用

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

作者： Zhenpeng Liang Hongjiao Guan Wenpeng Lu Xueping Peng Bing Xu Muyun Yang 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 Computing Power Internet and Service Computing Shandong Fundamental Research Center for Computer Science Jinan China Australian Artificial Intelligence Institute University of Technology Sydney Sydney Australia Faculty of Computing Harbin Institute of Technology Harbin China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The International Classification of Diseases (ICD) coding assigns standardized codes to diseases. Automating this process enhances the efficiency and accuracy of clinical records processing. However, current methods struggle with noisy and lengthy clinical texts, making it difficult to ensure the reliability of feature extraction. Furthermore, they typically make separate binary predictions for each code, overlooking the dependencies between them. To address these issues, we propose a novel model called Hierarchical Feature Optimization and Rescorer (HiRes). We employ a cascaded convolution architecture to mitigate noise and enhance feature representation. Additionally, a masked autoencoder-based rescorer is introduced to capture ICD code interdependencies, refining initial predictions for improved accuracy. The model also considers the inconsistencies in code representations. Experiments on the MIMIC datasets demonstrate the effectiveness of our model.

关键词： Codes Accuracy Computational modeling Noise Refining Encoding Reliability Speech processing Optimization Diseases

来源：评论

学校读者我要写书评

暂无评论

HiRes: Hierarchical Feature Optimization and Rescorer for Automatic ICD Coding

HiRes: Hierarchical Feature Optimization and Rescorer for Au...

引用

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

作者： Liang, Zhenpeng Guan, Hongjiao Lu, Wenpeng Peng, Xueping Xu, Bing Yang, Muyun 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 Computing Power Internet and Service Computing Shandong Fundamental Research Center for Computer Science Jinan China Australian Artificial Intelligence Institute University of Technology Sydney Sydney Australia Faculty of Computing Harbin Institute of Technology Harbin China

ISBN: (纸本)9798350368741

关键词： automatic ICD coding cascaded convolutional network multi-label classification

来源：评论

学校读者我要写书评

暂无评论

Learning with Open-world Noisy Data via Class-independent Margin in Dual Representation Space

arXiv

引用

arXiv 2025年

作者： Pan, Linchao Gao, Can Zhou, Jie Wang, Jinbao College of Computer Science and Software Engineering Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China

Learning with Noisy Labels (LNL) aims to improve the model generalization when facing data with noisy labels, and existing methods generally assume that noisy labels come from known classes, called closed-set noise. However, in real-world scenarios, noisy labels from similar unknown classes, i.e., open-set noise, may occur during the training and inference stage. Such open-world noisy labels may significantly impact the performance of LNL methods. In this study, we propose a novel dual-space joint learning method to robustly handle the open-world noise. To mitigate model overfitting on closed-set and open-set noises, a dual representation space is constructed by two networks. One is a projection network that learns shared representations in the prototype space, while the other is a One-Vs-All (OVA) network that makes predictions using unique semantic representations in the class-independent space. Then, bi-level contrastive learning and consistency regularization are introduced in two spaces to enhance the detection capability for data with unknown classes. To benefit from the memorization effects across different types of samples, class-independent margin criteria are designed for sample identification, which selects clean samples, weights closed-set noise, and filters open-set noise effectively. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods and achieves an average accuracy improvement of 4.55% and an AUROC improvement of 6.17% on CIFAR80N. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

TDMF: Text-Guided Denoising and Interactive Medical Image Fusion

TDMF: Text-Guided Denoising and Interactive Medical Image Fu...

引用

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

作者： Aimei Dong Jingyuan Xu Long Wang Guohua Lv Guixin Zhao Jinyong Cheng 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 Faculty of Computer Science and Technology Qilu University of Technology(Shandong Academy of Sciences) Shandong Provincial Key Laboratory of Computing Power Internet and Service Computing Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Multimodal image fusion aims to merge features from different modalities to create a comprehensively representative image. However, existing medical image fusion methods often struggle to handle noise generated during image acquisition, significantly diminishing their impact on visual quality. To address these challenges, we propose a semantically text-guided medical image fusion model, named TDMF. Specifically, TDMF guides classical image fusion through textual semantics and effectively coordinates the resolution of degradation and interaction issues during the fusion process. By integrating text encoders and interactive fusion modules, TDMF establishes a unified framework for denoising and interactive fusion of medical images. Extensive experiments have demonstrated that our proposed text-guided image fusion strategy offers significant advantages over state-of-the-art methods in medical image fusion performance.

关键词： Visualization Image resolution Semantics Noise reduction Noise Pipelines Noise measurement Speech processing Image fusion Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

A Federated Learning Network Intrusion Detection system for Multiple Imbalances

A Federated Learning Network Intrusion Detection System for ...

引用

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

作者： Yifan Zeng Peijia Zheng Jian Li School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Yunnan Key Laboratory of Service Computing Yunnan University of Finance and Economics Kunming China Key Laboratory of Computing Power Network and Information Security Ministry of Education Qilu University of Technology (Shandong Academy of Sciences) Jinan China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Centralized training of deep learning-based network intrusion detection systems (DLNIDS) raises privacy concerns and incurs huge overhead. Federated learning (FL), while preserving privacy, confronts challenges including size imbalance, heterogeneous and global imbalance. To address these issues, this paper proposes a Self-Balancing Federated learning Network Intrusion Detection system (SBFedNIDS) for accurate and efficient network intrusion detection while preserving privacy and tackling negative impact caused by multiple imbalances resulting from FL. SBFedNΠ)S can efficiently mitigate model bias cause by multiple imbalances and addresses imbalances at data-level via federated cost sensitive learning (FCSL) and client synthetic minority over-sampling technique (CSMOTE), without requiring additional training resources, unlike generative models. Extensive experiments on a benchmark dataset show that SBFedNIDS outperforms baselines in multiple imbalanced scenarios and even surpasses centralized training in several settings.

关键词： Training Privacy Data privacy Accuracy Costs Federated learning Network intrusion detection Signal processing Benchmark testing Speech processing

来源：评论

学校读者我要写书评

暂无评论

Identification of Non-Restorative Sleep Associated with Depression Using Ambulatory Electrocardiogram and Triaxial Acceleration

引用

IEEE Transactions on Biomedical Engineering 2025年 PP卷 PP页

作者： Liu, Li Wang, Chongyang Liu, Guangyuan Wen, Wanhui Southwest University Chongqing Key Laboratory of Generic Technology and System of Service Robots Chongqing Key Laboratory of Nonlinear Circuits and Intelligent Information Processing College of Electronic and Information Engineering Chongqing400715 China Sichuan University Key Laboratory of Rehabilitation Medicine in Sichuan Province Department of Rehabilitation Medicine Institute of Rehabilitation Medicine West China Hospital Chengdu610041 China

Objective: Non-restorative sleep is prevalent among individuals with depression and is strongly associated with the severity of the condition. Therefore, identifying non-restorative sleep can aid in the early screening of depression. Investigating non-restorative sleep in depression necessitates long-term monitoring under naturalistic conditions. Methods: In this study, we recruited 149 participants and collected electrocardiogram and triaxial acceleration from them, resulting in a total of 761 nights of data. The period from midnight to 6:30 AM was segmented into 78 five-minute intervals, from which 40 physiological features were extracted for each interval. To deal with variations in sleep patterns across participants and dates, we reordered the sleep data based on levels of parasympathetic nervous system (PNS) activation to explore the underlying neural mechanisms of non-restorative sleep in individuals with depressive symptoms. Results: We developed a model that integrated convolutional neural networks with an attention mechanism to identify nonrestorative sleep in individuals with depressive symptoms. The model demonstrated impressive performance on an independent test set, achieving an accuracy of 81.25% and an F1 score of 77.85%. Additionally, Bayes' theorem was used to compute the posterior probability indicating nonrestorative sleep in this population, assessing abnormal PNS activation. Conclusion: Finally, we designed a system capable of automatically evaluating nighttime sleep status and quantifying changes in non-restorative sleep associated with depression. Significance: This system offers a novel tool and method for the early identification of individuals at risk of depression. © 1964-2012 IEEE.

关键词： Sleep research

来源：评论

学校读者我要写书评

暂无评论

TDMF: Text-Guided Denoising and Interactive Medical Image Fusion

TDMF: Text-Guided Denoising and Interactive Medical Image Fu...

引用

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

作者： Dong, Aimei Xu, Jingyuan Wang, Long Lv, Guohua Zhao, Guixin Cheng, Jinyong Qilu University of Technology Shandong Academy of Sciences Jinan China Faculty of Computer Science and Technology Qilu University of Technology Shandong Academy of Sciences China Shandong Provincial Key Laboratory of Computing Power Internet and Service Computing Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (纸本)9798350368741

关键词： denoising Medical image fusion text-guided fusion

来源：评论

学校读者我要写书评

暂无评论

Unveiling the secrets without data: can graph neural networks be exploited through data-free model extraction attacks? 24

Unveiling the secrets without data: can graph neural network...

引用

Proceedings of the 33rd USENIX Conference on Security Symposium

作者： Yuanxin Zhuang Chuan Shi Mengmei Zhang Jinghui Chen Lingjuan Lyu Pan Zhou Lichao Sun Beijing University of Posts and Telecommunications Key Laboratory of Trustworthy Distributed Computing and Service (BUPT) Ministry of Education The Pennsylvania State University SONY AI Huazhong University of Science and Technology Lehigh University

ISBN: (纸本)9781939133441

Graph neural networks (GNNs) play a crucial role in various graph applications, such as social science, biology, and molecular chemistry. Despite their popularity, GNNs are still vulnerable to intellectual property threats. Previous studies have demonstrated the susceptibility of GNN models to model extraction attacks, where attackers steal the functionality of GNNs by sending queries and obtaining model responses. However, existing model extraction attacks often assume that the attacker has access to specific information about the victim model's training data, including node attributes, connections, and the shadow dataset. This assumption is impractical in real-world scenarios. To address this issue, we propose STEAL-GNN, the first data-free model extraction attack framework against GNNs. STEALGNN advances prior GNN extraction attacks in three key aspects: 1) It is completely data-free, as it does not require actual node features or graph structures to extract GNN models. 2) It constitutes a full-rank attack that can be applied to node classification and link prediction tasks, posing significant intellectual property threats across a wide range of graph applications. 3) It can handle the most challenging hard-label attack setting, where the attacker possesses no knowledge about the target GNN model and can only obtain predicted labels through querying the victim model. Our experimental results on four benchmark graph datasets demonstrate the effectiveness of STEALGNN in attacking representative GNN models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Impacts of terrain resolution on an extreme snowstorm process over the Southern Tibetan Plateau: Spatial distribution and cloud microphysical processes

引用

Atmospheric Research 2025年 322卷

作者： An, Pengchao Li, Ying Ye, Wei Fan, Xiaoting State Key Laboratory of Severe Weather Meteorological Science and Technology Chinese Academy of Meteorological Sciences Beijing China Department of Atmospheric and Oceanic Sciences Institute of Atmospheric Sciences Fudan University Shanghai China Ministry of Education Key Laboratory for Earth System Modeling Department of Earth System Science Tsinghua University Beijing China Sichuan Provincial Meteorological Service Centre Chengdu China

Precipitation is an essential component of the water cycle over the Tibetan Plateau (TP) in cold seasons. Investigating the impacts of various terrain resolutions in simulations on the precipitation distribution and rain-snow phase transition on the TP helps to understand the sources of simulation error and contributes to assessment and improvement. In this study, three types of terrain resolutions were used to conduct simulation experiments (namely WRF30s, WRF2m and WRF5m) on a large-scale extreme precipitation process over the TP under the influence of a storm over the Bay of Bengal in October 2008 near the Himalayas and the Yarlung Zangbo Grand Canyon, which serve as representative regions influenced by northward moisture. The results indicate that the terrain resolutions affect the precipitation amount and phase distribution. On the north of the Himalayas, a snowfall-dominated "phase bias" based on "wet bias" is observed above 3 km, which strengthens as the terrain resolution reduces, relating to the stronger water vapor flux transport and larger cold bias resulting from the terrain smoothing. The Bomi station, which is located in the canyon, undergoing a snow-to-rain phase transition influenced by the storm received further attention. The study not only shows the cloud microphysical characteristics of the precipitation on the TP under the influence of storms, but further reveals the snow melting phase transition weakening, manifested as the increase of snowfall accompanied by the decrease of terrain resolution. Further detailed analysis of cloud microphysical process in Bomi shows that the latter is related to the decrease melting rates of snow and graupel caused by the compression of the melting layer when the terrain near canyon increases with terrain resolution decrease. Consequently, the distribution and phase transition of precipitation on the TP induced by terrain resolutions requires attention during the simulation. © 2025 Elsevier B.V.

关键词： Cloud microphysical processes Extreme snowstorm Phase bias Terrain resolution The Tibetan Plateau Wet bias

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：