Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431523
Kaiming Shen;Ziping Zhao;Yannan Chen;Zepeng Zhang;Hei Victor Cheng
Fractional programming (FP) arises in various communications and signal processing problems because several key quantities in these fields are fractionally structured, e.g., the Cramér-Rao bound, the Fisher information, and the signal-to-interference-plus-noise ratio (SINR). A recently proposed method called the quadratic transform has been applied to the FP problems extensively. The main contributions of the present paper are two-fold. First, we investigate how fast the quadratic transform converges. To the best of our knowledge, this is the first work that analyzes the convergence rate for the quadratic transform as well as its special case the weighted minimum mean square error (WMMSE) algorithm. Second, we accelerate the existing quadratic transform via a novel use of Nesterov’s extrapolation scheme. Specifically, by generalizing the minorization-maximization (MM) approach, we establish a subtle connection between the quadratic transform and the gradient projection, thereby further incorporating the gradient extrapolation into the quadratic transform to make it converge more rapidly. Moreover, the paper showcases the practical use of the accelerated quadratic transform with two frontier wireless applications: integrated sensing and communications (ISAC) and massive multiple-input multiple-output (MIMO).
{"title":"Accelerating Quadratic Transform and WMMSE","authors":"Kaiming Shen;Ziping Zhao;Yannan Chen;Zepeng Zhang;Hei Victor Cheng","doi":"10.1109/JSAC.2024.3431523","DOIUrl":"10.1109/JSAC.2024.3431523","url":null,"abstract":"Fractional programming (FP) arises in various communications and signal processing problems because several key quantities in these fields are fractionally structured, e.g., the Cramér-Rao bound, the Fisher information, and the signal-to-interference-plus-noise ratio (SINR). A recently proposed method called the quadratic transform has been applied to the FP problems extensively. The main contributions of the present paper are two-fold. First, we investigate how fast the quadratic transform converges. To the best of our knowledge, this is the first work that analyzes the convergence rate for the quadratic transform as well as its special case the weighted minimum mean square error (WMMSE) algorithm. Second, we accelerate the existing quadratic transform via a novel use of Nesterov’s extrapolation scheme. Specifically, by generalizing the minorization-maximization (MM) approach, we establish a subtle connection between the quadratic transform and the gradient projection, thereby further incorporating the gradient extrapolation into the quadratic transform to make it converge more rapidly. Moreover, the paper showcases the practical use of the accelerated quadratic transform with two frontier wireless applications: integrated sensing and communications (ISAC) and massive multiple-input multiple-output (MIMO).","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3110-3124"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754864","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A fair beam allocation framework through reconfigurable intelligent surfaces (RISs) is proposed, incorporating the Max-min criterion. This framework focuses on designing explicit beamforming functionalities through optimization. Firstly, realistic models, grounded in geometrical optics, are introduced to characterize the input/output behaviors of RISs, effectively bridging the gap between the requirements on explicit beamforming operations and their practical implementations. Then, a highly efficient algorithm is developed for Max-min optimizations involving quadratic forms. Leveraging the Moreau-Yosida approximation, we successfully reformulate the original problem and propose an iterative algorithm to obtain the optimal solution. A comprehensive analysis of the algorithm’s convergence is provided. Importantly, this approach exhibits excellent extensibility, making it readily applicable to address a broader class of Max-min optimization problems. Finally, numerical and prototype experiments are conducted to validate the effectiveness of the framework. With the proposed beam allocation framework and algorithm, we clarify that several crucial redistribution functionalities of RISs, such as explicit beam-splitting, fair beam allocation, and wide-beam generation, can be effectively implemented. These explicit beamforming functionalities have not been thoroughly examined previously.
{"title":"Fair Beam Allocations Through Reconfigurable Intelligent Surfaces","authors":"Rujing Xiong;Ke Yin;Tiebin Mi;Jialong Lu;Kai Wan;Robert Caiming Qiu","doi":"10.1109/JSAC.2024.3431580","DOIUrl":"10.1109/JSAC.2024.3431580","url":null,"abstract":"A fair beam allocation framework through reconfigurable intelligent surfaces (RISs) is proposed, incorporating the Max-min criterion. This framework focuses on designing explicit beamforming functionalities through optimization. Firstly, realistic models, grounded in geometrical optics, are introduced to characterize the input/output behaviors of RISs, effectively bridging the gap between the requirements on explicit beamforming operations and their practical implementations. Then, a highly efficient algorithm is developed for Max-min optimizations involving quadratic forms. Leveraging the Moreau-Yosida approximation, we successfully reformulate the original problem and propose an iterative algorithm to obtain the optimal solution. A comprehensive analysis of the algorithm’s convergence is provided. Importantly, this approach exhibits excellent extensibility, making it readily applicable to address a broader class of Max-min optimization problems. Finally, numerical and prototype experiments are conducted to validate the effectiveness of the framework. With the proposed beam allocation framework and algorithm, we clarify that several crucial redistribution functionalities of RISs, such as explicit beam-splitting, fair beam allocation, and wide-beam generation, can be effectively implemented. These explicit beamforming functionalities have not been thoroughly examined previously.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3095-3109"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141755068","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431571
Nils Pachler;Edward F. Crawley;Bruce G. Cameron
With the reduced distance between satellites in modern megaconstellations, the potential for self-interference has emerged as a critical challenge that demands strategic solutions from satellite operators. The goal of this paper is to propose a cooperative framework that combines the Satellite Routing (i.e., mapping of beams to satellites) and Frequency Assignment (i.e., mapping of frequency spectrum to beams) strategies to mitigate self-interference both within and between satellites. This approach stands in contrast to current practices found in the literature, which address each problem independently and solely focus on intra-satellite interference. This study presents a novel methodology for addressing the Satellite Routing problem, specifically tailored for modern constellations to maximize capacity while effectively mitigating self-interference through the use of Integer Optimization. By combining this method with established Frequency Assignment techniques, the results demonstrate an increase in throughput of up to 138% for constellations such as SpaceX Starlink. Notably, the study reveals that relying on individual approaches to tackle interference may lead to undesired outcomes, underscoring the advantages of a cooperative framework. Through simulations, the study highlights the practicality and applicability of the proposed method under realistic operational conditions.
{"title":"Avoiding Self-Interference in Megaconstellations Through Cooperative Satellite Routing and Frequency Assignment","authors":"Nils Pachler;Edward F. Crawley;Bruce G. Cameron","doi":"10.1109/JSAC.2024.3431571","DOIUrl":"10.1109/JSAC.2024.3431571","url":null,"abstract":"With the reduced distance between satellites in modern megaconstellations, the potential for self-interference has emerged as a critical challenge that demands strategic solutions from satellite operators. The goal of this paper is to propose a cooperative framework that combines the Satellite Routing (i.e., mapping of beams to satellites) and Frequency Assignment (i.e., mapping of frequency spectrum to beams) strategies to mitigate self-interference both within and between satellites. This approach stands in contrast to current practices found in the literature, which address each problem independently and solely focus on intra-satellite interference. This study presents a novel methodology for addressing the Satellite Routing problem, specifically tailored for modern constellations to maximize capacity while effectively mitigating self-interference through the use of Integer Optimization. By combining this method with established Frequency Assignment techniques, the results demonstrate an increase in throughput of up to 138% for constellations such as SpaceX Starlink. Notably, the study reveals that relying on individual approaches to tackle interference may lead to undesired outcomes, underscoring the advantages of a cooperative framework. Through simulations, the study highlights the practicality and applicability of the proposed method under realistic operational conditions.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3188-3203"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754866","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431518
Sujunjie Sun;Weiwei Wu;Chenchen Fu;Xiaoxing Qiu;Junzhou Luo;Jianping Wang
This work studies the Age-of-Information (AoI) optimization problem in the information-gathering wireless network systems, where time-sensitive data updates are collected from multiple information sources, and each source is equipped with a battery and harvests energy from ambient energy, such as solar, wind, etc. The arrival of the harvested energy can be modeled as the stochastic process, and an information source can deliver its data update only when 1) there is energy in the battery, and 2) this source is selected to transmit its data update based on the transmission policy. This work analyzes how the energy arrival pattern of each source and the transmission policy jointly influence the average AoI among multiple sources. To the best of our knowledge, this is the first work that formally develops the closed-form expression of average AoI in the Stationary Randomized Sampling (SRS) policy space and proposes approximation schemes with constant ratios in multi-source systems under a stochastic energy harvesting model. More specifically, under the perfect wireless channel, the closed-form expression of AoI under the SRS policy space with arbitrary finite battery size is developed. Based on the result, we propose the Max Energy-Aware Weight (MEAW) policy, which is proven to achieve 2-approximation in the full policy space. Under the uncertain wireless channel, we develop the closed-form expression of Whittle’s index to address the target problem. Based on the result, we propose the Energy-aware Whittle’s index policy (EWIP) and prove its approximate performance by using the Lyapunov optimization techniques. Experimental results show that MEAW under the perfect channel setting and EWIP under the uncertain channel setting both perform close to the theoretical lower bound and outperform the state-of-the-art schemes.
{"title":"AoI Optimization in Multi-Source Update Network Systems Under Stochastic Energy Harvesting Model","authors":"Sujunjie Sun;Weiwei Wu;Chenchen Fu;Xiaoxing Qiu;Junzhou Luo;Jianping Wang","doi":"10.1109/JSAC.2024.3431518","DOIUrl":"10.1109/JSAC.2024.3431518","url":null,"abstract":"This work studies the Age-of-Information (AoI) optimization problem in the information-gathering wireless network systems, where time-sensitive data updates are collected from multiple information sources, and each source is equipped with a battery and harvests energy from ambient energy, such as solar, wind, etc. The arrival of the harvested energy can be modeled as the stochastic process, and an information source can deliver its data update only when 1) there is energy in the battery, and 2) this source is selected to transmit its data update based on the transmission policy. This work analyzes how the energy arrival pattern of each source and the transmission policy jointly influence the average AoI among multiple sources. To the best of our knowledge, this is the first work that formally develops the closed-form expression of average AoI in the Stationary Randomized Sampling (SRS) policy space and proposes approximation schemes with constant ratios in multi-source systems under a stochastic energy harvesting model. More specifically, under the perfect wireless channel, the closed-form expression of AoI under the SRS policy space with arbitrary finite battery size is developed. Based on the result, we propose the Max Energy-Aware Weight (MEAW) policy, which is proven to achieve 2-approximation in the full policy space. Under the uncertain wireless channel, we develop the closed-form expression of Whittle’s index to address the target problem. Based on the result, we propose the Energy-aware Whittle’s index policy (EWIP) and prove its approximate performance by using the Lyapunov optimization techniques. Experimental results show that MEAW under the perfect channel setting and EWIP under the uncertain channel setting both perform close to the theoretical lower bound and outperform the state-of-the-art schemes.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3172-3187"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754874","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431515
Shaoran Li;Nan Jiang;Yongce Chen;Weijun Xie;Wenjing Lou;Y. Thomas Hou
Channel State Information (CSI) is a critical piece of information for MU-MIMO beamforming. However, CSI estimation errors are inevitable in practice. The random and uncertain nature of CSI estimation errors poses significant challenges to MU-MIMO beamforming. State-of-the-art works addressing such a CSI uncertainty can be categorized into model-based and data-driven works, both of which have limitations when providing a performance guarantee to the users. In contrast, this paper presents Limited Sample-based Beamforming (LSBF)—a novel approach to MU-MIMO beamforming that only uses a limited number of CSI data samples (without assuming any knowledge of channel distributions). Thanks to the use of CSI data samples, LSBF enjoys flexibility similar to data-driven approaches and can provide a theoretical guarantee to the users—a major strength of model-based approaches. To achieve both, LSBF employs chance-constrained programming (CCP) and utilizes the $infty $ -Wasserstein ambiguity set to bridge the unknown CSI distribution with limited CSI samples. Through problem decomposition and a novel bilevel formulation for each subproblem based on limited CSI data samples, LSBF solves each subproblem with a binary search and convex approximation. We show that LSBF significantly improves the network performance while providing a probabilistic data rate guarantee to the users.
{"title":"MU-MIMO Beamforming With Limited Channel Data Samples","authors":"Shaoran Li;Nan Jiang;Yongce Chen;Weijun Xie;Wenjing Lou;Y. Thomas Hou","doi":"10.1109/JSAC.2024.3431515","DOIUrl":"10.1109/JSAC.2024.3431515","url":null,"abstract":"Channel State Information (CSI) is a critical piece of information for MU-MIMO beamforming. However, CSI estimation errors are inevitable in practice. The random and uncertain nature of CSI estimation errors poses significant challenges to MU-MIMO beamforming. State-of-the-art works addressing such a CSI uncertainty can be categorized into model-based and data-driven works, both of which have limitations when providing a performance guarantee to the users. In contrast, this paper presents Limited Sample-based Beamforming (LSBF)—a novel approach to MU-MIMO beamforming that only uses a limited number of CSI data samples (without assuming any knowledge of channel distributions). Thanks to the use of CSI data samples, LSBF enjoys flexibility similar to data-driven approaches and can provide a theoretical guarantee to the users—a major strength of model-based approaches. To achieve both, LSBF employs chance-constrained programming (CCP) and utilizes the \u0000<inline-formula> <tex-math>$infty $ </tex-math></inline-formula>\u0000-Wasserstein ambiguity set to bridge the unknown CSI distribution with limited CSI samples. Through problem decomposition and a novel bilevel formulation for each subproblem based on limited CSI data samples, LSBF solves each subproblem with a binary search and convex approximation. We show that LSBF significantly improves the network performance while providing a probabilistic data rate guarantee to the users.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3032-3047"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431579
Huayi Zhou;Ryan Seah;Marwan Jalaleddine;Warren J. Gross
Polar codes encounter challenges in decoder complexity while preserving good error-correction properties. Instead of conventional decoders, a quantum annealer (QA) decoder has been proposed to explore untapped possibilities. For future QA applications, a crucial prerequisite is transforming the optimization problem into quadratic unconstrained binary optimization (QUBO) form. However, existing QUBO forms for polar decoding result in suboptimal frame error rate (FER) performance for codes exceeding 8 bits. This paper redesigns the QUBO form for polar decoding. We first introduce a novel receiver constraint modeled by the binary cross-entropy (BCE) function. Utilizing a simulated annealing (SA) solver with the proposed QUBO form with BCE (QUBO-BCE) achieves maximum-likelihood (ML) performance for a code length of 32 bits. Next, to reduce the number of variables, we remove the frozen variables and introduce a simplified QUBO-BCE form (SQUBO-BCE). Additionally, CRC polynomials are modelled into constraints in QUBO form, resulting in a CRC-aided SQUBO-BCE (CA-SQUBO-BCE) form for polar decoding to further enhance the FER. Numerical results demonstrate that SQUBO-BCE achieves ML performance and reduces up to 61.5% of variables compared to QUBO-BCE. Furthermore, the proposed CA-SQUBO-BCE achieves near CRC-aided ML performance. The proposed SQUBO-BCE requires the lowest number of SA processes to reach a specific FER.
{"title":"Decoding of Polar Codes Using Quadratic Unconstrained Binary Optimization","authors":"Huayi Zhou;Ryan Seah;Marwan Jalaleddine;Warren J. Gross","doi":"10.1109/JSAC.2024.3431579","DOIUrl":"10.1109/JSAC.2024.3431579","url":null,"abstract":"Polar codes encounter challenges in decoder complexity while preserving good error-correction properties. Instead of conventional decoders, a quantum annealer (QA) decoder has been proposed to explore untapped possibilities. For future QA applications, a crucial prerequisite is transforming the optimization problem into quadratic unconstrained binary optimization (QUBO) form. However, existing QUBO forms for polar decoding result in suboptimal frame error rate (FER) performance for codes exceeding 8 bits. This paper redesigns the QUBO form for polar decoding. We first introduce a novel receiver constraint modeled by the binary cross-entropy (BCE) function. Utilizing a simulated annealing (SA) solver with the proposed QUBO form with BCE (QUBO-BCE) achieves maximum-likelihood (ML) performance for a code length of 32 bits. Next, to reduce the number of variables, we remove the frozen variables and introduce a simplified QUBO-BCE form (SQUBO-BCE). Additionally, CRC polynomials are modelled into constraints in QUBO form, resulting in a CRC-aided SQUBO-BCE (CA-SQUBO-BCE) form for polar decoding to further enhance the FER. Numerical results demonstrate that SQUBO-BCE achieves ML performance and reduces up to 61.5% of variables compared to QUBO-BCE. Furthermore, the proposed CA-SQUBO-BCE achieves near CRC-aided ML performance. The proposed SQUBO-BCE requires the lowest number of SA processes to reach a specific FER.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3204-3216"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431521
Tao Yu;Pengbo Huang;Shunqing Zhang;Xiaojing Chen;Yanzan Sun;Xin Wang
In order to provide design guidelines for energy efficient 6G networks, we propose a novel radial basis function (RBF) based optimization framework to maximize the integrated relative energy efficiency (IREE) metric. Different from the conventional energy efficient optimization schemes, we maximize the transformed utility for any given IREE using spectrum efficiency oriented RBF network and gradually update the IREE metric using proposed Dinkelbach’s algorithm. The existence and uniqueness properties of RBF networks are provided, and the convergence conditions of the entire framework are discussed as well. Through some numerical experiments, we show that the proposed IREE outperforms many existing SE or EE oriented designs and find a new Jensen-Shannon (JS) divergence constrained region, which behaves differently from the conventional EE-SE region. Meanwhile, by studying IREE-SE trade-offs under different traffic requirements, we suggest that network operators shall spend more efforts to balance the distributions of traffic demands and network capacities in order to improve the IREE performance, especially when the spatial variations of the traffic distribution are significant.
{"title":"IREE Oriented Green 6G Networks: A Radial Basis Function-Based Approach","authors":"Tao Yu;Pengbo Huang;Shunqing Zhang;Xiaojing Chen;Yanzan Sun;Xin Wang","doi":"10.1109/JSAC.2024.3431521","DOIUrl":"10.1109/JSAC.2024.3431521","url":null,"abstract":"In order to provide design guidelines for energy efficient 6G networks, we propose a novel radial basis function (RBF) based optimization framework to maximize the integrated relative energy efficiency (IREE) metric. Different from the conventional energy efficient optimization schemes, we maximize the transformed utility for any given IREE using spectrum efficiency oriented RBF network and gradually update the IREE metric using proposed Dinkelbach’s algorithm. The existence and uniqueness properties of RBF networks are provided, and the convergence conditions of the entire framework are discussed as well. Through some numerical experiments, we show that the proposed IREE outperforms many existing SE or EE oriented designs and find a new Jensen-Shannon (JS) divergence constrained region, which behaves differently from the conventional EE-SE region. Meanwhile, by studying IREE-SE trade-offs under different traffic requirements, we suggest that network operators shall spend more efforts to balance the distributions of traffic demands and network capacities in order to improve the IREE performance, especially when the spatial variations of the traffic distribution are significant.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3246-3261"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431568
Xiaoxin Su;Yipeng Zhou;Laizhong Cui;Quan Z. Sheng;Yinggui Wang;Song Guo
Federated learning (FL) has been extensively exploited in the training of machine learning models to preserve data privacy. In particular, wireless FL enables multiple clients to collaboratively train models by sharing model updates via wireless communication without exposing raw data. The state-of-the-art wireless FL advocates efficient aggregation of model updates from multiple clients by over-the-air computing. However, a significant deficiency of over-the-air aggregation lies in the infeasibility of TopK model compression given that top model updates cannot be aggregated directly before they are aligned according to their indices. In view of the fact that TopK can greatly accelerate FL, we design a novel wireless FL with voting based TopK algorithm, namely WFL-VTopK, so that top model updates can be aggregated by over-the-air computing directly. Specifically, there are two phases in WFL-VTopK. In Phase 1, clients vote their top model updates, based on which global top model updates can be efficiently identified. In Phase 2, clients formally upload global top model updates so that they can be directly aggregated by over-the-air computing. Furthermore, the convergence of WFL-VTopK is theoretically guaranteed under non-convex loss. Based on the convergence of WFL-VTopK, we optimize model utility subjecting to training time and energy constraints. To validate the superiority of WFL-VTopK, we extensively conduct experiments with real datasets under wireless communication. The experimental results demonstrate that WFL-VTopK can effectively aggregate models by only communicating 1%-2% top models updates, and hence significantly outperforms the state-of-the-art baselines. By significantly reducing the wireless communication traffic, our work paves the road to train large models in wireless FL.
{"title":"Fast-Convergent Wireless Federated Learning: A Voting-Based TopK Model Compression Approach","authors":"Xiaoxin Su;Yipeng Zhou;Laizhong Cui;Quan Z. Sheng;Yinggui Wang;Song Guo","doi":"10.1109/JSAC.2024.3431568","DOIUrl":"10.1109/JSAC.2024.3431568","url":null,"abstract":"Federated learning (FL) has been extensively exploited in the training of machine learning models to preserve data privacy. In particular, wireless FL enables multiple clients to collaboratively train models by sharing model updates via wireless communication without exposing raw data. The state-of-the-art wireless FL advocates efficient aggregation of model updates from multiple clients by over-the-air computing. However, a significant deficiency of over-the-air aggregation lies in the infeasibility of TopK model compression given that top model updates cannot be aggregated directly before they are aligned according to their indices. In view of the fact that TopK can greatly accelerate FL, we design a novel wireless FL with voting based TopK algorithm, namely WFL-VTopK, so that top model updates can be aggregated by over-the-air computing directly. Specifically, there are two phases in WFL-VTopK. In Phase 1, clients vote their top model updates, based on which global top model updates can be efficiently identified. In Phase 2, clients formally upload global top model updates so that they can be directly aggregated by over-the-air computing. Furthermore, the convergence of WFL-VTopK is theoretically guaranteed under non-convex loss. Based on the convergence of WFL-VTopK, we optimize model utility subjecting to training time and energy constraints. To validate the superiority of WFL-VTopK, we extensively conduct experiments with real datasets under wireless communication. The experimental results demonstrate that WFL-VTopK can effectively aggregate models by only communicating 1%-2% top models updates, and hence significantly outperforms the state-of-the-art baselines. By significantly reducing the wireless communication traffic, our work paves the road to train large models in wireless FL.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3048-3063"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-22DOI: 10.1109/JSAC.2024.3431516
Pengchao Han;Xingyan Shi;Jianwei Huang
Knowledge distillation (KD) can enable collaborative learning among distributed clients that have different model architectures and do not share their local data and model parameters with others. Each client updates its local model using the average model output/feature of all client models as the target, known as federated KD. However, existing federated KD methods often do not perform well when clients’ local models are trained with heterogeneous local datasets. In this paper, we propose Federated knowledge distillation enabled by Adversarial Learning ( FedAL ) to address the data heterogeneity among clients. First, to alleviate the local model output divergence across clients caused by data heterogeneity, the server acts as a discriminator to guide clients’ local model training to achieve consensus model outputs among clients through a min-max game between clients and the discriminator. Moreover, catastrophic forgetting may happen during the clients’ local training and global knowledge transfer due to clients’ heterogeneous local data. Towards this challenge, we design the less-forgetting regularization for both local training and global knowledge transfer to guarantee clients’ ability to transfer/learn knowledge to/from others. Experimental results show that FedAL and its variants achieve higher accuracy than other federated KD baselines.
{"title":"FedAL: Black-Box Federated Knowledge Distillation Enabled by Adversarial Learning","authors":"Pengchao Han;Xingyan Shi;Jianwei Huang","doi":"10.1109/JSAC.2024.3431516","DOIUrl":"10.1109/JSAC.2024.3431516","url":null,"abstract":"Knowledge distillation (KD) can enable collaborative learning among distributed clients that have different model architectures and do not share their local data and model parameters with others. Each client updates its local model using the average model output/feature of all client models as the target, known as federated KD. However, existing federated KD methods often do not perform well when clients’ local models are trained with heterogeneous local datasets. In this paper, we propose Federated knowledge distillation enabled by Adversarial Learning (\u0000<monospace>FedAL</monospace>\u0000) to address the data heterogeneity among clients. First, to alleviate the local model output divergence across clients caused by data heterogeneity, the server acts as a discriminator to guide clients’ local model training to achieve consensus model outputs among clients through a min-max game between clients and the discriminator. Moreover, catastrophic forgetting may happen during the clients’ local training and global knowledge transfer due to clients’ heterogeneous local data. Towards this challenge, we design the less-forgetting regularization for both local training and global knowledge transfer to guarantee clients’ ability to transfer/learn knowledge to/from others. Experimental results show that \u0000<monospace>FedAL</monospace>\u0000 and its variants achieve higher accuracy than other federated KD baselines.","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 11","pages":"3064-3077"},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141754879","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-17DOI: 10.1109/JSAC.2024.3412408
{"title":"IEEE Communications Society Information","authors":"","doi":"10.1109/JSAC.2024.3412408","DOIUrl":"https://doi.org/10.1109/JSAC.2024.3412408","url":null,"abstract":"","PeriodicalId":73294,"journal":{"name":"IEEE journal on selected areas in communications : a publication of the IEEE Communications Society","volume":"42 8","pages":"C3-C3"},"PeriodicalIF":0.0,"publicationDate":"2024-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10601512","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141725546","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：