检索结果-内蒙古大学图书馆

2024 International Conference on Computer Application and Information Security, ICCAIS 2024

作者： Hu, Junzhu Lu, Ying Xu, Quanyuan Li, Caiyi Wang, Mingyang Zhou, Hao Yang, Peidong Jiang, Feng College of Big Data and Intelligent Engineering Southwest Forestry University Kunming 650024 China Key Laboratory of Forestry and Ecological Big Data State Forestry Administration on Southwest Forestry University Kunming 650024 China Qujing Tobacco Company Qujing 655000 China

ISBN: (纸本)9781510689312

To address the challenges of vehicle attribute recognition in traffic management, intelligent parking systems, and road surveillance, this paper proposes an algorithm that integrates the Squeeze-and-Excitation (SE) attention mechanism into ResNet34. The SE mechanism is added after each BasicBlock, with a total of 16 SE modules incorporated to train the network. The optimal model is achieved after 100 iterations. The SE mechanism enhances the network’s ability to focus on key features by adaptively adjusting the importance of features across channels, thereby improving the effectiveness of feature representation. This dynamic adjustment allows the model to better adapt to complex and varying image structures, demonstrating strong adaptability in vehicle attribute recognition tasks. Experimental results show a significant improvement in recognition accuracy, with the enhanced ResNet34 model achieving a 5.1% increase in vehicle attribute recognition accuracy compared to the original ResNet34. The proposed algorithm effectively improves recognition performance, offering valuable technical support for the application of intelligent transportation systems. © The Authors.

关键词： ResNet34 SE attention mechanism vehicle attribute recognition

来源：评论

学校读者我要写书评

暂无评论

Improving Visual Reinforcement Learning via Spatial-Temporal data Augmentation

Improving Visual Reinforcement Learning via Spatial-Temporal...

引用

Consumer Electronics and Computer engineering (ICCECE), 2021 IEEE International Conference on

作者： Jiabao Li Qingwen Hou Jie Zhang Xianzhong Chen Key Laboratory of Knowledge Automation for Industrial Processes of Ministry of Education School of Automation and Electrical Engineering University of Science and Technology Beijing Beijing China

ISBN: (数字)9798331533113

ISBN: (纸本)9798331533120

Sample efficiency has been a challenging problem in visual reinforcement learning, where agents not only learn polices but also extract meaningful state representations for decision making from images. Reinforcement Learning methods that adopt data augmentation have significantly improved sample efficiency. However, prior studies have primarily focused on spatial augmentation, which enhances data diversity but insufficiently exploits temporal information, such as historical states which is critical for time-series-dependent reinforcement learning tasks. To address this gap, we introduce a Spatial-Temporal data Augmentation (STDA) framework designed to enhance the agent's ability to learn task-relevant state representations by effectively leveraging historical state information. We evaluate our approach using the DeepMind Control Suite, and empirical results demonstrate that STDA significantly improves performance, surpassing state-of-the-art methods across a range of tasks.

关键词： Visualization Q-learning Law enforcement Decision making Training data Reinforcement learning Benchmark testing data augmentation Consumer electronics

来源：评论

学校读者我要写书评

暂无评论

Counterfactual Inference for Generalized Zero-Shot Compound-Fault Diagnosis

引用

IEEE Transactions on Instrumentation and Measurement 2025年 74卷

作者： Xu, Juan Kong, Hui Ding, Xu Yuan, Xiaohui Hefei University of Technology Key Laboratory of Knowledge Engineering With Big Data Ministry of Education School of Computer and Information Hefei230601 China University of North Texas Department of Computer Science and Engineering DentonTX76205 United States

Learning a model heavily depends on the training examples, which are sometimes difficult to obtain if not impossible. This a typically true for fault diagnosis in machinery, particularly for compound faults. The counterfactual inference reveals the causal components inherent in the fault data in an interpretable manner, divulging critical causes from the observable phenomena. This article proposes a method to address the imbalance and interpretability issues of generalized zero-shot learning (GZSL) methods for compound-fault diagnosis using counterfactual inference. Our method uses a structural causal model (SCM) to decouple and generate fault features, which enhances the capabilities of the variational autoencoder and generative adversarial network (VAE-GAN) through a strengthened discriminator, and reveals the intrinsic causal components in fault data, distinguishing key fault causes from accompanying phenomena. This enables the classification of both single and compound faults by learning from examples of single faults, easing the dependence on the examples of compound faults. Extensive experimental results show that our method, trained solely with single-fault samples, achieves a harmonic average of 87.40% accuracy for both single and compound faults, outperforming existing state-of-the-art methods. This significantly improves both the accuracy and interpretability of compound-fault diagnosis. © 1963-2012 IEEE.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

SRv6-INT Enabled Network Monitoring and Measurement: Toward High-Yield Network Observability for Digital Twin 3rd

SRv6-INT Enabled Network Monitoring and Measurement: Toward...

引用

3rd International Conference on Machine Learning, Cloud Computing and Intelligent Mining, MLCCIM 2024

作者： Zhang, Xu Feng, Chuan Han, Pengchao Liu, Wei Gong, Xiaoxue Zhang, Qiupei Guo, Lei School of Communications and Information Engineering Chongqing University of Posts and Telecommunications Chongqing400065 China Institute of Intelligent Communication and Network Security Chongqing University of Posts and Telecommunications Chongqing400065 China Key Laboratory of Big Data Intelligent Computing Chongqing University of Posts and Telecommunications Chongqing400065 China School of Information Engineering Guangdong University of Technology Guangzhou510006 China

ISBN: (纸本)9789819616978

Network monitoring and measurement is an important part of realizing the network digital twin. However, it introduces the problem of high cost when obtaining the status data of physical networks. Therefore, to efficiently achieve network observability, we first introduce potential techniques and network scenarios for implementing network measurements. Then, a High-Yield Segment Routing over IPv6 based In-band Network Telemetry (HY-SRv6-INT) system is proposed in this paper. In particular, we design the HY-SRv6-INT system architecture. It consists of IPv6 switches implemented by the P4 language and a Software-Defined Networking (SDN) controller containing the HY-SRv6-INT application. Next, we experimentally demonstrate a packet-level path-associated network measurement process. In addition, the performance of the HY-SRv6-INT system is also quantitatively evaluated to verify its advantages. Finally, we discuss some research challenges and open issues for future network telemetry. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Internet protocols

来源：评论

学校读者我要写书评

暂无评论

Offline Reinforcement Learning via Conservative Smoothing and Dynamics Controlling

Offline Reinforcement Learning via Conservative Smoothing an...

引用

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

作者： Haihong Guo Fengxin Li Jiao Li Hongyan Liu School of Information Renmin University of China China Chinese Academy of Medical Sciences / Peking Union Medical College Institute of Medical Information / Medical Library China Key Laboratory of Data Engineering and Knowledge Engineering Ministry of Education China School of Economics and Management Tsinghua University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Offline Reinforcement Learning (RL) optimizes policy using pre-collected data instead of direct environment interaction, offering a safe and cost-effective solution for sequential decision-making in the real world. However, it faces challenges such as distribution shift issues and vulnerability under perturbations. Researchers have developed various conservative methods to improve the robustness of offline RL. Nevertheless, model-based methods can result in transition distribution shift issues, while model-free value-based uncertainty penalty methods may not be sufficiently robust. To address these problems, we propose a new method called Robust Offline RL via Conservative Smoothing and Dynamics Controlling (RCSD). To achieve reliable value estimation of out-of-distribution (OOD) actions, RCSD uses both model-free uncertainty penalty and model-based simulation methods. It introduces a new one-step simulation method with conservative dynamics controlling to avoid value overestimation caused by transition distribution shifts. Moreover, RCSD considers both current and next states when generating OOD states to ensure cautious value estimation and efficient data utilization. RCSD uses conservative Q-smoothing and policy smoothing to strengthen the policy against sudden changes under perturbations. Experiments on D4RL benchmark demonstrate that RCSD can achieve state-of-the-art performance compared to baselines in either benchmark or adversarial attack tests.

关键词： Uncertainty Smoothing methods Perturbation methods Estimation Reinforcement learning Benchmark testing Signal processing Robustness Speech processing Faces

来源：评论

学校读者我要写书评

暂无评论

A Model Inference Attack Based on Random Sampling in DLaaS 24th

A Model Inference Attack Based on Random Sampling in DLaaS

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Wu, Feng Sun, Shouyue Yang, Jiaxun Wu, Liwen Cui, Lei Qu, Youyang Yao, Shaowen Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China Engineering Research Center of Cyberspace Yunnan University Kunming China School of Information Science and Engineering Yunnan University Kunming China Data61 Commonwealth Scientific and Industrial Research Organization BurwoodVIC3166 Australia

ISBN: (纸本)9789819615506

Deep learning as a service (DLaaS) has become an effective business solution in numerous domains. Nevertheless, its service form has been proven vulnerable to model inference by previous research. Existing work on model inference attack relies on strong assumptions, such as the attacker knows the sample variance and model structure, which may not thoroughly reflect the attack’s potential threats. In this paper, we propose MC-infer, a zero-knowledge, real-data-free, and black-box model inference method inspired by Monte-Carlo sampling. In particular, MC-infer feeds random noises obtained from different distributions to the target model, and sniffs the corresponding target distributions according to its feedback. Then many samples are taken in these distributions to obtain enough robust noises. Finally, the distribution represented by the target model is fitted through these noises to perform model inference. Our extensive evaluations demonstrate that MC-infer can effectively infer the target model with less information, and general noise perturbation on the model’s outputs cannot defend against MC-infer. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Intelligent systems

来源：评论

学校读者我要写书评

暂无评论

KAN v.s. MLP for Offline Reinforcement Learning

KAN v.s. MLP for Offline Reinforcement Learning

引用

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

作者： Haihong Guo Fengxin Li Jiao Li Hongyan Liu School of Information Renmin University of China China Institute of Medical Information / Medical Library Chinese Academy of Medical Sciences / Peking Union Medical College China Key Laboratory of Data Engineering and Knowledge Engineering Ministry of Education China School of Economics and Management Tsinghua University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Kolmogorov-Arnold Networks (KAN) is an emerging neural network architecture in machine learning. It has greatly interested the research community about whether KAN can be a promising alternative to the commonly used Multi-Layer Perceptions (MLP). Experiments in various fields demonstrated that KAN-based machine learning can achieve comparable if not better performance than MLP-based methods, but with much smaller parameter scales and are more explainable. In this paper, we explore the incorporation of KAN into the actor and critic networks for offline reinforcement learning (RL). We evaluated the performance, parameter scales, and training efficiency of various KAN and MLP-based conservative Q-learning (CQL) on the classical D4RL benchmark for offline RL. Our study demonstrates that KAN can achieve performance close to the commonly used MLP with significantly fewer parameters. This allows us to choose the base networks according to the offline RL task requirements.

关键词： Training Performance evaluation Hands Q-learning Neural networks Memory management Signal processing Benchmark testing Splines (mathematics) Speech processing

来源：评论

学校读者我要写书评

暂无评论

TraSculptor: Visual Analytics for Enhanced Decision-Making in Road Traffic Planning

引用

IEEE Transactions on Visualization and Computer Graphics 2025年 PP卷 PP页

作者： Deng, Zikun Liu, Yuanbang Zhu, Mingrui Xiang, Da Yu, Haiyue Su, Zicheng Lu, Qinglong Schreck, Tobias Cai, Yi South China University of Technology School of Software Engineering China South China University of Technology Ministry of Education Key Laboratory of Big Data and Intelligent Robot China Tongji University Key Laboratory of Road and Traffic Engineering of the Ministry of Education China National University of Singapore Department of Civil and Environmental Engineering Singapore Graz University of Technology Austria

The design of urban road networks significantly influences traffic conditions, underscoring the importance of informed traffic planning. Traffic planning experts rely on specialized platforms to simulate traffic systems, assessing the efficacy of the road network across various states of modifications. Nevertheless, a prevailing issue persists: many existing traffic planning platforms exhibit inefficiencies in flexibly interacting with the road network's structure and attributes and intuitively comparing multiple states during the iterative planning process. This paper introduces TraSculptor, an interactive planning decision-making system. To develop TraSculptor, we identify and address two challenges: interactive modification of road networks and intuitive comparison of multiple network states. For the first challenge, we establish flexible interactions to enable experts to easily and directly modify the road network on the map. For the second challenge, we design a comparison view with a history tree of multiple states and a road-state matrix to facilitate intuitive comparison of road network states. To evaluate TraSculptor, we provided a usage scenario where the Braess's paradox was showcased, invited experts to perform a case study on the Sioux Falls network, and collected expert feedback through interviews. © 1995-2012 IEEE.

关键词： Visual analytics

来源：评论

学校读者我要写书评

暂无评论

SecureCut: Federated Gradient Boosting Decision Trees with Efficient Machine Unlearning 27th

SecureCut: Federated Gradient Boosting Decision Trees with ...

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Li, Bowen Zhang, Jian Li, Jie Wu, Chentao Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China Yancheng Blockchain Research Institute Yancheng China Blockchain Advanced Research Center Shanghai China Shanghai Key Laboratory of Trusted Data Circulation and Governance and Web3 Shanghai China

ISBN: (纸本)9783031781681

In response to legislation, companies are now mandated to honor the right to be forgotten by erasing user data. Consequently, it has become imperative to enable data removal in Vertical Federated Learning (VFL), where multiple parties provide private features for model training. In VFL, data removal, i.e., machine unlearning, requires removing specific features across all samples, ensuring privacy in federated learning. To address this challenge, we propose SecureCut, a novel Gradient Boosting Decision Tree (GBDT) framework that effectively enables both instance unlearning and feature unlearning without the need for retraining from scratch. Leveraging a robust GBDT structure, we enable effective data deletion while reducing degradation of model performance. Extensive experimental results on popular datasets demonstrate that our method achieves superior model utility and forgetfulness compared to the state-of-the-art methods. To the best of our knowledge, this is the first work that investigates machine unlearning in VFL scenarios. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

End-to-end robust and transparent image watermarking method for screen-shooting scenarios

引用

Journal of Electronic Imaging 2025年第2期34卷

作者： Yang, Nana Chen, Yi Wang, Hongxia Shen, Guowei Guo, Chun Cui, Yunhe Liu, Yong Wu, Hanzhou Guizhou University College of Computer Science and Technology State Key Laboratory of Public Big Data Guiyang China Guizhou University Guizhou Provincial Characteristic Key Laboratory of Software Engineering and Information Security Guiyang China Sichuan University School of Cyber Science and Engineering Chengdu China Hubei University of Automotive Technology School of Electrical and Information Engineering Shiyan China Shanghai University School of Communication and Information Engineering Shanghai China

Screen-shooting robust image watermarking techniques are the primary means for copyright protection and leakage traceability of screen content. However, existing image watermarking methods generate watermarked images with visible artifacts and are not robust enough against attacks. Therefore, we propose an end-to-end robust and transparent image watermarking method for protecting copyright in screen-shooting scenarios. In this method, we introduce scaling distortion to combine with color distortion, moiré distortion, Gaussian noise, and JPEG compression to design a new noise layer. In the noise layer, we propose two ways to simulate real moiré distortion. We also propose a training strategy adding five distortions in stages during end-to-end adversarial training. Both the noise layer and training strategy can help improve robustness. In addition, we propose an edge-aware loss function to guide the encoder to focus on modifying the edge regions of the cover image, thereby enhancing transparency. The experimental results show that, under different shooting conditions, the watermark extraction accuracy exceeds 97%, indicating the high robustness of our proposed method. In terms of transparency, the peak signal-to-noise ratio reaches 38.51 dB, which is 1.34 dB higher than the state-of-the-art deep learning-based screen-shooting watermarking method, PIMoG. © 2025 SPIE and IS&T.

关键词： edge-aware loss image watermarking noise layer screen-shooting training strategy

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：