检索结果-内蒙古大学图书馆

Reduced-RSS: Reduced Residual Search Space for Nash Bargaining Solutions

IEEE Transactions on Vehicular Technology 2025年第5期74卷 8215-8225页

作者： Cha, Chaeyeon Park, Hyunggon Ewha Womans University Graduate Program in Smart Factory Department of Electronic and Electrical Engineering Seoul03760 Korea Republic of

A fair and efficient resource allocation is crucial to maximize the efficiency of vehicular networks with time-varying resources among dynamically changing vehicles. The Nash Bargaining Solution (NBS) from the cooperative game theory has been one of the fair and optimal resource management solutions for multiple users. However, the computational complexity required to find the NBS for general utility functions exponentially increases either as the number of users increases or as the resources change over time. This is because the size of bargaining sets where the NBS should be found changes. In this paper, we propose an algorithm to reduce the size of the search spaces by considering the adjacent bargaining sets correlated by resource changes. We use the axiom of independence of linear transformations in NBS for the linear approximation of NBS in the adjacent bargaining sets based on its linear transform, leading to reduced residual search space. Moreover, we prove that the NBS is always included in the reduced residual search space based on the axiom of independence of irrelevant alternatives in NBS. Through simulations and experiments on synthetic and real traffic data, we demonstrate that the complexity of existing algorithms for NBS is lowered with the reduced residual search spaces. © 1967-2012 IEEE.

关键词： NIST Resource management Heuristic algorithms Vectors Dynamical systems Computational complexity Accuracy Vehicle dynamics Dynamic scheduling System performance

来源：评论

学校读者我要写书评

暂无评论

SNeRV: Spectra-preserving Neural Representation for Video

arXiv

引用

arXiv 2025年

作者： Kim, Jina Lee, Jihoo Kang, Je-Won Department of Electronic and Electrical Engineering Graduate Program in Smart Factory Ewha Womans University Seoul Korea Republic of

Neural representation for video (NeRV), which employs a neural network to parameterize video signals, introduces a novel methodology in video representations. However, existing NeRV-based methods have difficulty in capturing fine spatial details and motion patterns due to spectral bias, in which a neural network learns high-frequency (HF) components at a slower rate than low-frequency (LF) components. In this paper, we propose spectra-preserving NeRV (SNeRV) as a novel approach to enhance implicit video representations by efficiently handling various frequency components. SNeRV uses 2D discrete wavelet transform (DWT) to decompose video into LF and HF features, preserving spatial structures and directly addressing the spectral bias issue. To balance the compactness, we encode only the LF components, while HF components that include fine textures are generated by a decoder. Specialized modules, including a multi-resolution fusion unit (MFU) and a high-frequency restorer (HFR), are integrated into a backbone to facilitate the representation. Furthermore, we extend SNeRV to effectively capture temporal correlations between adjacent video frames, by casting the extension as additional frequency decomposition to a temporal domain. This approach allows us to embed spatio-temporal LF features into the network, using temporally extended up-sampling blocks (TUBs). Experimental results demonstrate that SNeRV outperforms existing NeRV models in capturing fine details and achieves enhanced reconstruction, making it a promising approach in the field of implicit video representations. The codes are available at https://***/qwertja/SNeRV. Copyright © 2025, The Authors. All rights reserved.

关键词： Discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Label Space-Induced Pseudo Label Refinement for Multi-Source Black-Box Domain Adaptation

引用

IEEE Transactions on Image Processing 2025年 34卷 3181-3193页

作者： Yoo, Chaehwa Liu, Xiaofeng Xing, Fangxu Woo, Jonghye Kang, Je-Won Chungbuk National University School of Electrical Engineering Cheongju Korea Republic of Massachusetts General Hospital Harvard Medical School Gordon Center for Medical Imaging Dept. of Radiology BostonMA United States Yale University New Haven United States Ewha W. University Department of Electronic and Electrical Engineering Graduate Program in Smart Factory Seoul Korea Republic of

Conventional unsupervised domain adaptation (UDA) requires access to source data and/or source model parameters, prohibiting its practical application in terms of privacy, security, and intellectual property. Recent black-box UDA (BDA) reduces such constraints by defining a pseudo label from a single encapsulated source application programming interface (API) prediction, which allows for self-training of the target model. Nonetheless, existing methods have limited consideration for multi-source settings, in which multiple source domain APIs are available to generate pseudo labels. In this work, we introduce a novel training framework for multi-source BDA (MSBDA), dubbed Label Space-Induced Pseudo Label Refinement (LPR). Specifically, LPR incorporates a Pseudo label Refinery Network (PRN) that learns the relationship among source domains conditioned by the target domain only utilizing source API’s prediction. The target model is adapted by our dual phases PRN. First, a warm-up phase targets to avoid failure due to noisy samples and provide an initial pseudo-label, which is followed by a label refinement phase with domain relationship exploration. We provide theoretical support for the mechanism of the LPR. Experimental results on four benchmark datasets demonstrate that MSBDA using LPR achieves competitive performance compared to state-of-the-art approaches with different DA settings. © 1992-2012 IEEE.

关键词： Adaptation models Labeling Data models Predictive models Training Closed box Application programming interfaces Upper bound Sports Smoothing methods

来源：评论

学校读者我要写书评

暂无评论

Neural Volumetric Video Coding With Hierarchical Coded Representation of Dynamic Volume

引用

IEEE Transactions on Multimedia 2025年

作者： Shin, Ju-Yeon Lee, Jung-Kyung Bang, Gun Kim, Jun-Sik Kang, Je-Won Ewha W. University Department of Electronic and Electrical Engineering and Graduate Program in Smart Factory Seoul03760 Korea Republic of Ewha W. University Department of Electronic and Electrical Engineering Korea Republic of Ofinno United States Media Coding Research Lab. Daejeon34129 Korea Republic of

This paper proposes a novel multi-view (MV) video coding technique that leverages a four-dimensional (4D) voxel-grid representation to enhance coding efficiency, particularly in novel view synthesis. Although the voxel grid approximation provides a continuous representation for dynamic scenes, its volumetric nature requires substantial storage. The compression of MV videos can be interpreted as the compression of dense features. However, the substantial size of these features poses a significant problem relative to the generation of dynamic scenes at arbitrary viewpoints. To address this challenge, this study introduces a hierarchical coded representation of dynamic volumes based on low-rank tensor decomposition of volumetric features and develops effective coding techniques based on this representation. The proposed method employs a two-level coding strategy to capture the temporal characteristics of the decomposed features. At a higher level, spatial features are encoded, representing 3D structural information, with time-invariant components over short intervals of an MV video sequence. At a lower level, temporal features are encoded to capture the dynamics of current scenes. The spatial features are shared in a group, and temporal features are encoded at each time step. The experimental results demonstrate that the proposed technique outperforms existing MV video coding standards and current state-of-the-art methods, providing superior rate-distortion performance in the novel view synthesis of MV video compression. © 1999-2012 IEEE.

关键词： Electric distortion

来源：评论

学校读者我要写书评

暂无评论

An Agent-Based Model of Microgrids with Distributed Renewable Energy Sources

An Agent-Based Model of Microgrids with Distributed Renewabl...

引用

Power, Energy and electrical engineering (CPEEE), International Conference on

作者： Gyu Yeon Kim Danyao Peng Sangyub Lee Seon Han Choi Graduate Program in Smart Factory Ewha Womans University Seoul South Korea Department of Electronic and Electrical Engineering Ewha Womans University Seoul South Korea Energy IT Convergence Research Center Korea Electronics Technology Institute Seongnam South Korea

ISBN: (数字)9798331520090

ISBN: (纸本)9798331520106

The high penetration of renewable energy sources poses a challenge to stability of power systems, particularly microgrids, which are more vulnerable due to their small capacity and low inertia. To address instability, system components and their interactions need to be represented to more efficiently understand and optimize complex systems. Agent-based modeling enables autonomous operation and collaborative interaction, making the system adaptable to changes efficiently, making it ideal for microgrid management. In this paper, an agent-based microgrid simulation model is proposed. By independently representing microgrid components such as generators, energy storage systems (ESS), loads, environments and controllers, models can dynamically simulate the behavior of power systems, respond to system changes in real time. To validate the proposed model, experiments are conducted based on IEEE 13 bus system, and the results show that with the use of ESS, the frequency peak is reduced from over 62 Hz to around 60 Hz. In the case of sudden demand fluctuations, the frequency remains relatively stable, staying close to 60 Hz. The model demonstrates flexibility and robustness in managing renewable energy fluctuations and sudden load changes.

关键词： Renewable energy sources Adaptation models Fluctuations Microgrids Power system stability Stability analysis Real-time systems smart grids Thermal stability Load modeling

来源：评论

学校读者我要写书评

暂无评论

Low-complexity Anomaly Detection Method based on Feature Importance using Shapley Value 14

Low-complexity Anomaly Detection Method based on Feature Imp...

引用

14th International Conference on Ubiquitous and Future Networks, ICUFN 2023

作者： Rheey, Joohong Park, Hyunggon Ewha Womans University Graduate Program in Smart Factory Dept. Electronic and Electrical Engineering Seoul Korea Republic of

ISBN: (纸本)9798350335385

The increasing popularity of Internet of Things (IoT) devices has brought significant security challenges to IoT networks. However, most deep learning-based anomaly detection solutions often require high computation performance so that it is difficult to be implanted on low-end IoT devices with limited power and memory capacity. In this paper, we propose a low-complexity network anomaly detection method based on feature selection using the Shapley value for the Isolation Forest algorithm. The proposed feature selection method using the Shapley value can reduce the dimension of input data, thereby improving the performance with reduced computational complexity. We provide simulation results to demonstrate the effectiveness of the proposed method. The results show that the proposed method based on Isolation Forest achieves comparable performance to the deep learning method based on neural networks while using fewer dimensions than the deep learning method. © 2023 IEEE.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Service-Based Optimal Group Resource Allocation Strategy 14

Service-Based Optimal Group Resource Allocation Strategy

引用

14th International Conference on Ubiquitous and Future Networks, ICUFN 2023

作者： Cha, Chaeyeon Park, Hyunggon Ewha Womans University Graduate Program in Smart Factory Dept. Electronic and Electrical Engineering Seoul Korea Republic of

ISBN: (纸本)9798350335385

With the increasing development of the Internet of Things (IoT), there is a growing need for efficient resource allocation to the many devices in the IoT system. This need is particularly acute in the case of time-sensitive applications, where fast resource allocation is essential. However, the resource allocation based on the Nash bargaining solution (NBS) necessitates exponentially increasing computational complexity for user dynamic systems. We propose an approach that makes groups based on the requested services and predicts a disagreement point whenever the group size changes, with the aim of reducing the computational complexity to find the NBS. Through simulation results, we demonstrate that our proposed approach outperforms the existing method in terms of execution time to find the NBS, maintaining the optimality in finding NBS. © 2023 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Learning feature alignment and dual correlation for few‐shot image classification

引用

CAAI Transactions on Intelligence Technology 2024年第2期9卷 303-318页

作者： Xilang Huang Seon Han Choi Department of Electronic and Electrical Engineering Ewha Womans UniversitySeoulSouth Korea Graduate Program in Smart Factory Ewha Womans UniversitySeoulSouth Korea

Few‐shot image classification is the task of classifying novel classes using extremely limited labelled *** perform classification using the limited samples,one solution is to learn the feature alignment(FA)information between the labelled and unlabelled sample *** FA methods use the feature mean as the class prototype and calculate the correlation between prototype and unlabelled features to learn an alignment ***,mean prototypes tend to degenerate informative features because spatial features at the same position may not be equally important for the final classification,leading to inaccurate correlation ***,the authors propose an effective intraclass FA strategy that aggregates semantically similar spatial features from an adaptive reference prototype in low‐dimensional feature space to obtain an informative prototype feature map for precise correlation ***,a dual correlation module to learn the hard and soft correlations was developed by the *** module combines the correlation information between the prototype and unlabelled features in both the original and learnable feature spaces,aiming to produce a comprehensive cross‐correlation between the prototypes and unlabelled *** both FA and cross‐attention modules,our model can maintain informative class features and capture important shared features for *** results on three few‐shot classification benchmarks show that the proposed method outperformed related methods and resulted in a 3%performance boost in the 1‐shot setting by inserting the proposed module into the related methods.

关键词： image classification machine learning metric learning

来源：评论

学校读者我要写书评

暂无评论

An Improved I-V model for Resistive Random Access Memory

An Improved I-V model for Resistive Random Access Memory

引用

2022 IEEE International Conference on electrical, Computer, and Energy Technologies, ICECET 2022

作者： Chung, Harry Shin, Hyungsoon Sun, Wookyung Park, Jisun Graduate Program in Smart Factory Ewha Womans University Dept. of Electronic and Electrical Engineering Seoul Korea Republic of Seoul National University Department of Electrical and Computer Engineering Seoul Korea Republic of Graduate Program in Smart Factory Ewha Womans University Seoul Korea Republic of

ISBN: (数字)9781665470872

ISBN: (纸本)9781665470872

In this paper, we performed a simulation for a physical resistive random-access device (RRAM) to propose an improved model representing its electrical characteristics. Voltage ThrEshold Adaptive Memristor (VTEAM) model that operates based on the voltage threshold of the device adapted for the RRAM simulation. We applied some physical features to the model to adjust the internal state variable to improve the simulation results. The internal state variable used in the model was designed after the physical feature such as the distance between the end of the filament and the electrode of the device. Therefore, applying some of these physical conditions affects the simulation results. Simulation results with the newly applied model showed optimized I-V characteristics that resemble the physical device's behavior more than the existing model. © 2022 IEEE.

关键词： Memristors

来源：评论

学校读者我要写书评

暂无评论

Adaptive Loss Function Design Algorithm for Input Data Distribution in Autoencoder 13

Adaptive Loss Function Design Algorithm for Input Data Distr...

引用

13th International Conference on Information and Communication Technology Convergence, ICTC 2022

作者： Rheey, Joohong Choi, Dayoung Park, Hyunggon Smart Factory Multidisciplinary Program Korea Republic of Ewha Womans University Dept. Electronic and Electrical Engineering Seoul Korea Republic of

ISBN: (数字)9781665499392

ISBN: (纸本)9781665499392

The training performance of an autoencoder is significantly affected by its loss function. In order to improve the performance of autoencoders, it is important to design ap-propriate loss functions. However, many loss functions have been designed without taking into account the characteristics of the input data distribution. In this paper, we propose an algorithm for the design of a loss function by adaptively determining optimal parameters to input data distribution. Specifically, the proposed optimal parameters for loss function are determined by explicitly considering the dependency of the standard deviation of input data distribution. The simulation results confirm that the pro-posed algorithm can improve traditional loss functions. Moreover, the loss function determined by the proposed algorithm can reduce the computational complexity for training autoencoders. © 2022 IEEE.

关键词： Learning systems

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：