检索结果-内蒙古大学图书馆

Mean Field-Based Dynamic Backoff Optimization for MIMO-Enabled Grant-Free NOMA inMassive IoT Networks

Journal on Internet of Things 2024年第1期6卷 17-41页

作者： Haibo Wang Hongwei Gao Pai Jiang Matthieu De Mari Panzer Gu Yinsheng Liu School of Electronic and Information Engineering Beijing Jiaotong UniversityBeijing100044China Department of Information Systems Technology and Design Pillar Singapore University of Technology and DesignSingapore487372Singapore Nokia Group Alcatel Lucent Shanghai BellShanghai200120China

In the 6G Internet of Things(IoT)paradigm,unprecedented challenges will be raised to provide massive connectivity,ultra-low latency,and energy efficiency for ultra-dense IoT *** address these challenges,we explore the non-orthogonal multiple access(NOMA)based grant-free random access(GFRA)schemes in the cellular uplink to support massive IoT devices with high spectrum efficiency and low access *** particular,we focus on optimizing the backoff strategy of each device when transmitting time-sensitive data samples to a multiple-input multiple-output(MIMO)-enabled base station subject to energy *** cope with the dynamic varied channel and the severe uplink interference due to the uncoordinated grant-free access,we formulate the optimization problem as a multi-user non-cooperative dynamic stochastic game(MUN-DSG).To avoid dimensional disaster as the device number grows large,the optimization problem is transformed into a mean field game(MFG),and its Nash equilibrium can be achieved by solving the corresponding Hamilton-Jacobi-Bellman(HJB)and Fokker-Planck-Kolmogorov(FPK)***,a Mean Field-based Dynamic Backoff(MFDB)scheme is proposed as the optimal GFRA solution for each *** simulation has been fulfilled to compare the proposed MFDB with contemporary random access approaches like access class barring(ACB),slotted-Additive LinksOn-lineHawaii Area(ALOHA),andminimum backoff(MB)under both static and dynamic channels,and the results proved thatMFDB can achieve the least access delay and cumulated cost during multiple transmission frames.

关键词： 6G internet of things grant-free random access NOMA dynamic backoff mechanism mean field game

来源：评论

学校读者我要写书评

暂无评论

Testing Group Fairness via Optimal Transport Projections 38

Testing Group Fairness via Optimal Transport Projections

引用

38th International Conference on Machine Learning, ICML 2021

作者： Si, Nian Murthy, Karthyek Blanchet, Jose Nguyen, Viet Anh Department of Management Science & Engineering Stanford University United States Engineering Systems and Design Pillar Singapore University of Technology and Design Singapore VinAI Research Viet Nam

ISBN: (纸本)9781713845065

We present a statistical testing framework to detect if a given machine learning classifier fails to satisfy a wide range of group fairness notions. The proposed test is a flexible, interpretable, and statistically rigorous tool for auditing whether exhibited biases are intrinsic to the algorithm or due to the randomness in the data. The statistical challenges, which may arise from multiple impact criteria that define group fairness and which are discontinuous on model parameters, are conveniently tackled by projecting the empirical measure onto the set of group-fair probability models using optimal transport. This statistic is efficiently computed using linear programming and its asymptotic distribution is explicitly obtained. The proposed framework can also be used to test for testing composite fairness hypotheses and fairness with multiple sensitive attributes. The optimal transport testing formulation improves interpretability by characterizing the minimal covariate perturbations that eliminate the bias observed in the audit. Copyright © 2021 by the author(s)

关键词： Linear programming

来源：评论

学校读者我要写书评

暂无评论

Adaptive Federated Learning via New Entropy Approach

arXiv

引用

arXiv 2023年

作者： Zheng, Shensheng Yuan, Wenhao Wang, Xuehe Duan, Lingjie The School of Artificial Intelligence Sun Yat-sen University Zhuhai519082 China The Pillar of Engineering Systems and Design Singapore University of Technology and Design 467372 Singapore

Federated Learning (FL) has emerged as a prominent distributed machine learning framework that enables geographically discrete clients to train a global model collaboratively while preserving their privacy-sensitive data. However, due to the non-independent-and-identically-distributed (Non-IID) data generated by heterogeneous clients, the performances of the conventional federated optimization schemes such as FedAvg and its variants deteriorate, requiring the design to adaptively adjust specific model parameters to alleviate the negative influence of heterogeneity. In this paper, by leveraging entropy as a new metric for assessing the degree of system disorder, we propose an adaptive FEDerated learning algorithm based on ENTropy theory (FedEnt) to alleviate the parameter deviation among heterogeneous clients and achieve fast convergence. Nevertheless, given the data disparity and parameter deviation of heterogeneous clients, determining the optimal dynamic learning rate for each client becomes a challenging task as there is no communication among participating clients during the local training epochs. To enable a decentralized learning rate for each participating client, we first introduce the mean-field terms to estimate the components associated with other clients’ local parameters. Furthermore, we provide rigorous theoretical analysis on the existence and determination of the mean-field estimators. Based on the mean-field estimators, the closed-form adaptive learning rate for each client is derived by constructing the Hamilton equation. Moreover, the convergence rate of our proposed FedEnt is proved. The extensive experimental results on the real-world datasets (i.e., MNIST, EMNIST-L, CIFAR10, and CIFAR100) show that our FedEnt algorithm surpasses FedAvg and its variants (i.e., FedAdam, FedProx, and FedDyn) under Non-IID settings and achieves a faster convergence rate. Copyright © 2023, The Authors. All rights reserved.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Exploiting Constructive Interference Precoding for MISO-NOMA Networks

Exploiting Constructive Interference Precoding for MISO-NOMA...

引用

IEEE International Conference on Communications (ICC)

作者： Wei Wang Lingjie Duan Xin Liu Nan Zhao School of Information and Communication Engineering Dalian University of Technology Dalian P. R. China Engineering Systems and Design Pillar Singapore University of Technology and Design (SUTD) Singapore

As a symbol-level precoding scheme, constructive interference precoding (CIP) has been demonstrated its superiority in multi-antenna orthogonal multiple access (OMA) systems. By utilizing both the channel state information (CSI) and data symbols, harmful multi-user interference can convert to useful reception power via the well-designed CIP precoder. When CIP meets non-orthogonal multiple access (NOMA) whose bottleneck is usually at the weaker user side, this paper is the first to propose CIP to enhance the downlink MISO-NOMA networks, by making the desired signal of the stronger user in a typical NOMA pair constructive to the weaker user. In our CIP-NOMA scheme, we properly design the CIP precoder for transmit power minimization at the base station (BS), subject to signal-to-interference-plus-noise ratio (SINR) requirements of NOMA users. By replacing the non-convex constraints with convex linear matrix inequalities, we optimally obtain the precoding solutions for our CIP-NOMA scheme by semi-definite relaxation (SDR) method. Moreover, as compared to conventional NOMA schemes, we theoretically prove that once two NOMA users possess distinct channel gains, our optimized CIP-NOMA scheme always uses smaller transmit power to reach the target SINR thresholds. We further extend our CIP-NOMA scheme to the scenario of imperfect CSI, by further addressing the hidden CSI errors. Finally, we run extensive simulations to verify that our proposed CIP-NOMA scheme greatly outperforms zero-forcing (ZF), conventional NOMA and conventional CIP schemes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An End-to-End Bi-Objective Approach to Deep Graph Partitioning

SSRN

引用

SSRN 2024年

作者： Wei, Pengcheng Fang, Yuan Wen, Zhihao Xiao, Zheng Chen, Binbin Information Systems Technology and Design Pillar Singapore University of Technology and Design 485998 Singapore School of Computing and Information Systems Singapore Management University 178902 Singapore College of Computer Science and Electronic Engineering Hunan University Hunan410082 China

desirable to partition a large graph into smaller partitions. The bi-objective graph partitioning problem aims to minimize both the "cut" across partitions and the imbalance among the sizes of partitions. Existing heuristic methods either suffer scalability issues or ignore partition-wise balance, resulting in inferior partitioning quality. On the other hand, recent deep learning-based methods only utilize node-level features without paying attention to the broader features on the graph, and none of them is trained in a truly end-to-end framework, which leads poor performance on partitioning. In this paper, we propose a novel method based on graph neural networks (GNN) which employs multilevel graph features, and solves the problem end-to- end through a bi-objective formulation. Specifically, we explore node-, local-, and global-level graph features. Then, we propose a well-bounded bi-objective objective function which not only minimizes the cut, but focuses on partition-wise balance across all partitions. Finally, we propose a GNN-based deep model with a Hardmax operator, which enables the model to optimize partitions in a truly end-to-end manner. Experimental results on 11 datasets in various applications and scales show that our method can achieve notable improvements in partitioning quality and scalability than existing bi-objective baselines and deep graph partitioning baselines. © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Poster: Optimal Variance-Reduced Client Sampling for Multiple Models Federated Learning

Poster: Optimal Variance-Reduced Client Sampling for Multipl...

引用

International Conference on Distributed Computing systems

作者： Haoran Zhang Zekai Li Zejun Gong Marie Siew Carlee Joe-Wong Rachid El-Azouzi Electrical and Computer Engineering Carnegie Mellon University Pittsburgh PA USA Information Systems Technology and Design Pillar Singapore University of Technology and Design Singapore CERI/LIA University of Avignon Avignon France

ISBN: (数字)9798350386059

ISBN: (纸本)9798350386066

Federated learning (FL) is a variant of distributed learning in which multiple clients collaborate to learn a global model without sharing their data with the central server. In real-world scenarios, a client may be involved in training multiple unrelated FL models, which we call multi-model federated learning (MMFL), and the client sampling strategy and task allocation are crucial for improving system performance. In this paper, we propose an optimal sampling method to minimize the variance of global updates for unbiased learning in MMFL systems. The resulting method achieves an average accuracy of over 30 % higher than other baseline methods, as we demonstrate through simulations on real-world federated datasets.

关键词： Training Accuracy Federated learning Computational modeling System performance Distributed databases Sampling methods

来源：评论

学校读者我要写书评

暂无评论

Positive and unlabeled learning with controlled probability boundary fence 24

Positive and unlabeled learning with controlled probability ...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Changchun Li Yuanchao Dai Lei Feng Ximing Li Bing Wang Jihong Ouyang College of Computer Science and Technology and Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University China Information Systems Technology and Design Pillar Singapore University of Technology and Design Singapore

Positive and Unlabeled (PU) learning refers to a special case of binary classification, and technically, it aims to induce a binary classifier from a few labeled positive training instances and loads of unlabeled instances. In this paper, we derive a theorem indicating that the probability boundary of the asymmetric disambiguation-free expected risk of PU learning is controlled by its asymmetric penalty, and we further empirically evaluated this theorem. Inspired by the theorem and its empirical evaluations, we propose an easy-to-implement two-stage PU learning method, namely Positive and Unlabeled Learning with Controlled Probability Boundary Fence (PUL-CPBF). In the first stage, we train a set of weak binary classifiers concerning different probability boundaries by minimizing the asymmetric disambiguation-free empirical risks with specific asymmetric penalty values. We can interpret these induced weak binary classifiers as a probability boundary fence. For each unlabeled instance, we can use the predictions to locate its class posterior probability and generate a stochastic label. In the second stage, we train a strong binary classifier over labeled positive training instances and all unlabeled instances with stochastic labels in a self-training manner. Extensive empirical results demonstrate that PUL-CPBF can achieve competitive performance compared with the existing PU learning baselines.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Adaptive Resource Allocation in Quantum Key Distribution (QKD) for Federated Learning

Adaptive Resource Allocation in Quantum Key Distribution (QK...

引用

International Conference on Computing, Networking and Communications (ICNC)

作者： Rakpong Kaewpuang Minrui Xu Dusit Niyato Han Yu Zehui Xiong Xuemin Sherman Shen School of Computer Science and Engineering Nanyang Technological University (NTU) Pillar of Information Systems Technology and Design Singapore University of Technology and Design (SUTD) Department of Electrical and Computer Engineering University of Waterloo

Increasing privacy and security concerns in intelligence-native 6G networks require quantum key distribution-secured federated learning (QKD-FL), in which data owners connected via quantum channels can train an FL global model collaboratively without exposing their local datasets. To facilitate QKD-FL, the architectural design and routing management framework are essential. However, effective implementation is still lacking. To this end, we propose a hierarchical architecture for QKD-FL systems in which QKD resources (i.e., wavelengths) and routing are jointly optimized for FL applications. In particular, we focus on adaptive QKD resource allocation and routing for FL workers to minimize the deployment cost of QKD nodes under various uncertainties, including security requirements. The experimental results show that the proposed architecture and the resource allocation and routing model can reduce the deployment cost by 7.72% compared to the CO-QBN algorithm.

关键词： Adaptation models Costs Uncertainty Adaptive systems Federated learning Computer architecture Routing

来源：评论

学校读者我要写书评

暂无评论

CT3D++: Improving 3D Object Detection with Keypoint-induced Channel-wise Transformer

arXiv

引用

arXiv 2024年

作者： Sheng, Hualian Cai, Sijia Zhao, Na Deng, Bing Liang, Qiao Zhao, Min-Jian Ye, Jieping Alibaba Cloud Hangzhou310052 China Information Systems Technology and Design Pillar Singapore University of Technology and Design 487372 Singapore College of Information Science and Electronic Engineering Zhejiang University Hangzhou310027 China

The field of 3D object detection from point clouds is rapidly advancing in computer vision, aiming to accurately and efficiently detect and localize objects in three-dimensional space. Current 3D detectors commonly fall short in terms of flexibility and scalability, with ample room for advancements in performance. In this paper, our objective is to address these limitations by introducing two frameworks for 3D object detection with minimal hand-crafted design. Firstly, we propose CT3D, which sequentially performs raw-point-based embedding, a standard Transformer encoder, and a channel-wise decoder for point features within each proposal. Secondly, we present an enhanced network called CT3D++, which incorporates geometric and semantic fusion-based embedding to extract more valuable and comprehensive proposal-aware information. Additionally, CT3D ++ utilizes a point-to-key bidirectional encoder for more efficient feature encoding with reduced computational cost. By replacing the corresponding components of CT3D with these novel modules, CT3D++ achieves state-of-the-art performance on both the KITTI dataset and the large-scale Waymo Open Dataset. The source code for our frameworks will be made accessible at https://***/hlsheng1/CT3D-plusplus. Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Joint Resource Pricing and Quality Control for Cloud Mining Services in Blockchain Networks: A Game-Theoretic Analysis

Joint Resource Pricing and Quality Control for Cloud Mining ...

引用

IEEE Conference on Global Communications (GLOBECOM)

作者： Licheng Ye Xian Xiu Zehui Xiong Lin Gao School of Electronics and Information Engineering and the Guangdong Provincial Key Laboratory of Aerospace Communication and Networking Technology Harbin Institute of Technology Shenzhen China Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore

ISBN: (数字)9798350351255

ISBN: (纸本)9798350351262

The mining process in blockchain networks generates substantial computing consumption, which can be very challenging for miners operated by mobile users (MUs) with limited computing resources. Cloud mining service (CMS) offers a viable solution to this challenge by allowing miners to offload computation-intensive mining tasks to cloud mining providers (CMP) with abundant resources. One key problem in such a scenario is to design an effective resource pricing mechanism for the CMP. Existing researches in this field mainly focused on the differentiated pricing mechanisms, where different MUs are required to pay different prices, which are often very complicated. In this work, we explore an efficient resource pricing mechanism, where the CMP sets a uniform price for all MUs but regulates the quality of resources for different MUs. Such a mechanism can capture the key essence of differentiated pricing and greatly reduce complexity. Based on this novel mechanism, we establish a two-stage Stackelberg game between the CMP and MUs, which consists of a resource pricing and quality control problem (for the CMP) as the first stage and a mining competition game (among all MUs) as the second stage. We derive the closed-form Nash equilibrium for the mining competition game in the second stage and propose a successive convex approximation (SCA)-based algorithm that converges to the near-optimal solution in the first stage. Simulation results show that the proposed iterative algorithm can improve system utility by 32.5% compared to other schemes.

关键词： Simulation Pricing Games Quality control Nash equilibrium Blockchains Complexity theory Iterative methods Global communication Mobile computing

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：