检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Han, Haochen Zheng, Qinghua Luo, Minnan Miao, Kaiyao Tian, Feng Chen, Yan The National Engineering Lab for Big Data Analytics Xi’an Jiaotong University Xi’an710049 China The School of Computer Science and Technology Xi’an Jiaotong University Xi’an710049 China The Key Laboratory of Intelligent Networks and Network Security Xi’an Jiaotong University Ministry of Education Xi’an710049 China The School of Cyber Science and Engineering Xi’an Jiaotong University Xi’an710049 China

Recently, video recognition is emerging with the help of multi-modal learning, which focuses on integrating distinct modalities to improve the performance or robustness of the model. Although various multi-modal learning methods have been proposed and offer remarkable recognition results, almost all of these methods rely on high-quality manual annotations and assume that modalities among multi-modal data provide semantically relevant information. Unfortunately, the widely used video datasets are usually coarse-annotated or collected from the Internet. Thus, it inevitably contains a portion of noisy labels and noisy correspondence. To address this challenge, we use the audio-visual action recognition task as a proxy and propose a noise-tolerant learning framework to find anti-interference model parameters against both noisy labels and noisy correspondence. Specifically, our method consists of two phases that aim to rectify noise by the inherent correlation between modalities. First, a noise-tolerant contrastive training phase is performed to make the model immune to the possible noisy-labeled data. Despite the benefits brought by contrastive training, it would overfit the noisy correspondence and thus provide false supervision. To alleviate the influence of noisy correspondence, we propose a cross-modal noise estimation component to adjust the consistency between different modalities. As the noisy correspondence existed at the instance level, we further propose a category-level contrastive loss to reduce its interference. Second, in the hybrid-supervised training phase, we calculate the distance metric among features to obtain corrected labels, which are used as complementary supervision to guide the training. Furthermore, due to the lack of suitable datasets, we establish a benchmark of real-world noisy correspondence in audio-visual data by relabeling the Kinetics dataset. Extensive experiments on a wide range of noisy levels demonstrate that our method significant

关键词： Modal analysis

来源：评论

学校读者我要写书评

暂无评论

Transformer Model with Multi-Type Classification Decisions for Intrusion Attack Detection of Track Traffic and Vehicle

Transformer Model with Multi-Type Classification Decisions f...

引用

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

作者： Quanlong Guan Tian Zhang Yu Qin Yuyu Zhou Yangguang Zhu Yuansheng Zhong Xiujie Huang Zhifei Duan Zhefu Li Changjiang Liu Xiaofeng Wu College of Information Science and Technology Jinan University Guangzhou China Guangdong Key Laboratory of Data Security and Privacy Preserving China School of Intelligent Systems Science and Engineering Jinan University Zhuhai China Guangdong Testing Institute of Product Quality Supervision Guangdong China Guangzhou XPeng Motors Technology Co. Ltd Guangdong China Guangzhou Polytechnic of Sports Guangzhou China

security vulnerabilities, illustrated by the menace of track traffic or vehicle hacking, present a substantial risk to the Controller Area Network (CAN) bus, enabling unauthorized remote access and intrusion. Nevertheless, existing vehicle intrusion detection models encounter challenges in capturing temporal aspects, compromising the preservation of temporal attributes in the data. Addressing this predicament, we propose a novel Vehicle Intrusion Detection system (PTIDS) model based on Principal Component Analysis (PCA) and Transformer architecture, equipped with multi-type classification decisions. This model utilizes the PCA algorithm to preprocess the data and employs a multi-head self-attention mechanism to simulate the continuous input of the real-world environment in vehicle or track traffic data. Experimental results demonstrate that the PTIDS model outperforms traditional neural networks regarding accuracy, precision, recall, and F1-score for vehicle intrusion detection, confirming the importance of temporal variables in intrusion detection. However, owing to the high accuracy and low discrimination of the PTIDS model on the Car Hacking dataset, we migrated the model to the M-CAN and B-CAN intrusion datasets and conducted ablation experiments. The results reveal that the model achieved an accuracy of 90.14% on the M-CAN dataset, surpassing other models by 54.5%, demonstrating robust generalization ability.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Size-invariance matters: rethinking metrics and losses for imbalanced multi-object salient object detection 24

Size-invariance matters: rethinking metrics and losses for i...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Feiran Li Qianqian Xu Shilong Bao Zhiyong Yang Runmin Cong Xiaochun Cao Qingming Huang Institute of Information Engineering Chinese Academy of Sciences Beijing China and School of Cyber Security University of Chinese Academy of Sciences Beijing China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China Institute of Information Science Beijing Jiaotong University Beijing China and School of Control Science and Engineering Shandong University Jinan China and Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Cyber Science and Tech. Shenzhen Campus Sun Yat-sen University School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China and Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China and Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China

This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. In pursuit of this, we propose a generic approach that evaluates each salient object separately and then combines the results, effectively alleviating the imbalance. We further develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes. Theoretically, we provide evidence supporting the validity of our new metrics and present the generalization analysis of SOD. Extensive experiments demonstrate the effectiveness of our method. The code is available at https://***/Ferry-Li/SI-SOD.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Structural disorder-induced topological phase transitions in quasicrystals

arXiv

引用

arXiv 2024年

作者： Peng, Tan Xiong, Yong-Chen Hua, Chun-Bo Liu, Zheng-Rong Zhu, Xiaolu Cao, Wei Lv, Fang Hou, Yue Zhou, Bin Wang, Ziyu Xiong, Rui Key Laboratory of Artificial Micro- and Nano-Structures of Ministry of Education School of Physics and Technology Wuhan University Wuhan430072 China Hubei Key Laboratory of Energy Storage and Power Battery School of Mathematics Physics and Optoelectronic Engineering Hubei University of Automotive Technology Shiyan442002 China School of Electronic and Information Engineering Hubei University of Science and Technology Xianning437100 China Department of Physics Hubei University Wuhan430062 China The Institute of Technological Science Wuhan University Wuhan430072 China Key Laboratory of Intelligent Sensing System and Security of Ministry of Education Hubei University Wuhan430062 China

Recently, the structural disorder-induced topological phase transitions in periodic systems have attracted much attention. However, in aperiodic systems such as quasicrystalline systems, the interplay between structural disorder and band topology is still unclear. In this work, we investigate the effects of structural disorder on a quantum spin Hall insulator phase and a higher-order topological phase in a two-dimensional Amman-Beenker tiling quasicrystalline lattice, respectively. We demonstrate that the structural disorder can induce a topological phase transition from a quasicrystalline normal insulator phase to an amorphous quantum spin Hall insulator phase, which is confirmed by bulk gap closing and reopening, robust edge states, quantized spin Bott index and conductance. Furthermore, the structural disorder-induced higher-order topological phase transition from a quasicrystalline normal insulator phase to an amorphous higher-order topological phase characterized by quantized quadrupole moment and topological corner states is also found. More strikingly, the disorder-induced higher-order topological insulator with eight corner states represents a distinctive topological state that eludes realization in conventional crystalline systems. Our work extends the study of the interplay between disorder effects and topologies to quasicrystalline and amorphous systems. Copyright © 2024, The Authors. All rights reserved.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

Deep Joint Source-Channel Coding for CSI Feedback: An End-to-End Approach

arXiv

引用

arXiv 2022年

作者： Xu, Jialong Ai, Bo Wang, Ning Chen, Wei State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University Beijing100044 China State Key Laboratory of Rail Traffic Control and Safety The School of Electronic and Information Engineering The Frontiers Science Center for Smart High-Speed Railway System Beijing Jiaotong University Beijing100044 China Research Center of Networks and Communications Peng Cheng Laboratory Shenzhen518055 China Henan Joint International Research Laboratory of Intelligent Networking and Data Analysis Zhengzhou University Zhengzhou450001 China School of Information Engineering Zhengzhou University Zhengzhou450001 China

The increased throughput brought by MIMO technology relies on the knowledge of channel state information (CSI) acquired in the base station (BS). To make the CSI feedback overhead affordable for the evolution of MIMO technology (e.g., massive MIMO and ultra-massive MIMO), deep learning (DL) is introduced to deal with the CSI compression task. Based on the separation principle in existing communication systems, DL based CSI compression is used as source coding. However, this separate source-channel coding (SSCC) scheme is inferior to the joint source-channel coding (JSCC) scheme in the finite blocklength regime. In this paper, we propose a deep joint source-channel coding (DJSCC) based framework for the CSI feedback task. In particular, the proposed method can simultaneously learn from the CSI source and the wireless channel. Instead of truncating CSI via Fourier transform in the delay domain in existing methods, we apply non-linear transform networks to compress the CSI. Furthermore, we adopt an SNR adaption mechanism to deal with the wireless channel variations. The extensive experiments demonstrate the validity, adaptability, and generality of the proposed framework. © 2022, CC BY.

关键词： Channel state information

来源：评论

学校读者我要写书评

暂无评论

Size-Invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection

arXiv

引用

arXiv 2024年

作者： Li, Feiran Xu, Qianqian Bao, Shilong Yang, Zhiyong Cong, Runmin Cao, Xiaochun Huang, Qingming Institute of Information Engineering Chinese Academy of Sciences Beijing China School of Cyber Security University of Chinese Academy of Sciences Beijing China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China Institute of Information Science Beijing Jiaotong University Beijing China School of Control Science and Engineering Shandong University Jinan China Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Cyber Science and Tech. Sun Yat-Sen University Shenzhen Campus China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Image Encryption Methods in Deep Joint Source Channel Coding: A Review and Performance Evaluation

Image Encryption Methods in Deep Joint Source Channel Coding...

引用

International Conference on Computer and Communications (ICCC)

作者： Jialong Xu Bo Ai Wei Chen Ang Yang Peng Sun State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University Beijing China Frontiers Science Center for Smart High-speed Railway System Beijing China Research Center of Networks and Communications Peng Cheng Laboratory Shenzhen China Henan Joint International Research Laboratory of Intelligent Networking and Data Analysis Zhengzhou University Zhengzhou China Communication Research Institute vivo Mobile Communication Co. Ltd. Beijing China

ISBN: (纸本)9781665409513

With the introduction of the deep learning (DL), joint source and channel coding (JSCC) has achieved great success and has been demonstrated to have better performance especially in the case of low signal-to-noise ratios (SNRs) and limited bandwidth. As a promising wireless transmission technology, DL based JSCC (DJSCC) can serve different kinds of applications. However, some applications (e.g., biomedical, surveillance, financial and military) have strict requirements on data privacy and data security. Traditional image encryption methods are based on the shuffle operation and changing the intensity of each pixel in the plain image, which changes the ambient structure of the picture and degrades the reconstruct performance of the DJSCC. Perceptual image encryption methods for deep neural networks (DNNs) have been demonstrated to have similar classification accuracy in comparison to the DNNs trained on plain images. However, it is unknown how to conduct image encryption methods in the DJSCC framework. In this paper, we investigate different image encryption methods at first and compare their reconstruction performance. This research work provide a guidance for selecting the image encryption method suitable for DJSCC transmission.

关键词： Deep learning Wireless communication Performance evaluation data privacy Surveillance Neural networks Encryption

来源：评论

学校读者我要写书评

暂无评论

Prototype Implementation and Experimental Evaluation for LoRa-Backscatter Communication systems with RF Energy Harvesting and Low Power Management

引用

IEEE Transactions on Communications 2024年

作者： Tang, Xiaoqing Liu, Xin Xie, Guihui Cui, Yongqiang Li, Dong Hubei University School of Artificial Intelligence Wuhan China Ministry of Education Key Laboratory of Intelligent Sensing System and Security Hubei University Wuhan China Wuhan Qingchuan University Institute of Electronic Information Wuhan China School of Automation Wuhan China South-Central Minzu University College of Electronics and Information Engineering Wuhan China Macau University of Science and Technology School of Computer Science and Engineering Avenida Wai Long Taipa Macau China

Battery free Internet of Things (BF IoT), which is realized by harvesting ambient energy to power the IoT devices, has many advantages such as low power, low cost, free-maintenance, easy deployment, and environmental protection. As the two key enabling technologies of BF IoT, the integration of energy harvesting, management and backscatter communication (BackCom) is expected to meet practical applications. However, several fundamental issues need to be addressed to fully develop the potential of such integration. In this paper, we propose an ultra-low power energy harvesting and management scheme that effectively reduces the startup and quiescent power consumption. In the case of extremely limited energy harvested, the Long Range (LoRa) BackCom digital waveform generation algorithm with ultra-low power consumption and low complexity is studied, which can support low-power and low-cost micro-controllers (MCUs). Then, a radio frequency (RF) front-end scheme that is compatible with RF energy harvesting (RFEH), BackCom, and On-Off-keying (OOK) receiving is proposed. On this basis, we design and implement a fully functional BF LoRa Tag, which integrates tag antennas, LoRa BackCom, low-power OOK receiving, multi-parameter real-time sensing, RFEH and management, at a cost of only $6. Our test results show that the BF LoRa Tag can be self-powered by harvesting the RF energy as low as -19 dBm, with the cold-start power as low as 280 nW, the quiescent power as low as 49 nW, the OOK receiving power consumption as low as 1.74 μW, the LoRa BackCom power consumption of only 251 μW, and the communication distance up to 445 meters. Compared with the prototypes in existing literature, the BF LoRa Tag not only has lower power consumption, wider coverage, and the lowest cost, but also has complete system functions. Finally, we discuss the reciprocity between stations and receivers, as well as the BF LoRa cellular communication network, which provides useful references for the practica

关键词： Mobile telecommunication systems

来源：评论

学校读者我要写书评

暂无评论

Long and Short-Term Constraints Driven Safe Reinforcement Learning for Autonomous Driving

arXiv

引用

arXiv 2024年

作者： Hu, Xuemin Chen, Pan Wen, Yijun Tang, Bo Chen, Long School of Artificial Intelligence Hubei University Wuhan430062 China Key Laboratory of Intelligent Sensing System and Security Hubei University Ministry of Education Hubei Wuhan430062 China Department of Electrical and Computer Engineering Worcester Polytechnic Institute WorcesterMA01609 United States State Key Laboratory of Multimodal Artificial Intelligence Systems State Key Laboratory of Management and Control for Complex Systems Chinese Academy of Sciences Beijing100190 China WAYTOUS Inc. Beijing100083 China Shenzhen518107 China Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University Xi’an710049 China

Reinforcement learning (RL) has been widely used in decision-making and control tasks, but the risk is very high for the agent in the training process due to the requirements of interaction with the environment, which seriously limits its industrial applications such as autonomous driving systems. Safe RL methods are developed to handle this issue by constraining the expected safety violation costs as a training objective, but the occurring probability of an unsafe state is still high, which is unacceptable in autonomous driving tasks. Moreover, these methods are difficult to achieve a balance between the cost and return expectations, which leads to learning performance degradation for the algorithms. In this paper, we propose a novel algorithm based on the long and short-term constraints (LSTC) for safe RL. The short-term constraint aims to enhance the shortterm state safety that the vehicle explores, while the long-term constraint enhances the overall safety of the vehicle throughout the decision-making process, both of which are jointly used to enhance the vehicle safety in the training process. In addition, we develop a safe RL method with dual-constraint optimization based on the Lagrange multiplier to optimize the training process for end-to-end autonomous driving. Comprehensive experiments were conducted on the MetaDrive simulator. Experimental results demonstrate that the proposed method achieves higher safety in continuous state and action tasks, and exhibits higher exploration performance in long-distance decision-making tasks compared with state-of-the-art methods. Copyright © 2024, The Authors. All rights reserved.

关键词： Lagrange multipliers

来源：评论

学校读者我要写书评

暂无评论

Tensor Dynamic Fusion Based Modality-Imbalanced Multimodal Federated Learning in Mobile Edge Computing for Consumer Applications

引用

IEEE Transactions on Consumer Electronics 2024年

作者： Nie, Xin Yi, Lingzhi Yang, Laurence T. Deng, Xianjun Fan, Fulan Zhang, Hongtao Huazhong University of Science and Technology Hubei Key Laboratory of Distributed System Security Hubei EngineeringResearch Center on Big Data Security School of Cyber Science and Engineering Wuhan430074 China Zhongnan University of Economics and Law School of Information and Safety Engineering Wuhan430074 China Huazhong University of Science and Technology School of Computer Science and Technology Wuhan China Zhengzhou University School of Computer and Artificial Intelligence Zhengzhou China St. Francis Xavier University Department of Computer Science Antigonish Canada South-Central Minzu University School of Education Wuhan430074 China

Mobile edge computing extends the capabilities of computation and storage from the cloud to the edge, enabling local data storage and processing. This computing paradigm complements federated learning (FL), as it satisfies the need for privacy protection while providing efficiency and quick response. In consumer applications, multimodal data is experiencing explosive growth. Therefore, multimodal FL has emerged as a crucial research area. However, applying traditional FL methods directly to multimodal data poses certain challenges. Additionally, issues like sensor failures result in modality-imbalanced data among clients, i.e., the number and type of data modalities between client devices vary. Hence, it is imperative to design specialized FL frameworks and models specifically tailored for analyzing multimodal data. To address these challenges, we propose a tensor fusion-based multimodal federated learning (TenMFL). First, a tensor dynamic fusion network is designed for addressing the challenge of multimodal data fusion. This approach separates the parameter forward computation and the modal fusion computation, effectively resolving issues related to client-side modality imbalance. Second, we introduce a hybrid fusion method that combines feature fusion and decision fusion to further exploit the complementarity of multimodal data. We evaluate TenMFL on three multimodal datasets, experimental results demonstrate our superior performances. © 1975-2011 IEEE.

关键词： Federated learning

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：