Pub Date : 2025-07-22DOI: 10.1109/TMC.2025.3591822
Licheng Ye;Zehui Xiong;Lin Gao;Dusit Niyato
Mobile edge computing (MEC) is a promising technology that enhances the efficiency of mobile blockchain networks, by enabling miners, often acted by mobile users (MUs) with limited computing resources, to offload resource-intensive mining tasks to nearby edge computing servers. Collaborative block mining can further boost mining efficiency by allowing multiple miners to form coalitions, pooling their computing resources and transaction data together to mine new blocks collaboratively. Therefore, an MEC-assisted collaborative blockchain network can leverage the strengths of both technologies, offering improved efficiency, security, and scalability for blockchain systems. While existing research in this area has mainly focused on the single-coalition collaboration mode, where each miner can only join one coalition, this work explores a more comprehensive multi-coalition collaboration mode, which allows each miner to join multiple coalitions. To analyze the behavior of miners and the edge computing service provider (ECP) in this scenario, we propose a novel two-stage Stackelberg game. In Stage I, the ECP, as the leader, determines the prices of computing resources for all MUs. In Stage II, each MU decides the coalitions to join, resulting in an overlapping coalition formation (OCF) game; Subsequently, each coalition decides how many edge computing resources to purchase from the ECP, leading to an edge resource competition (ERC) game. We derive the closed-form Nash equilibrium for the ERC game, based on which we further propose an OCF-based alternating algorithm to achieve a stable coalition structure for the OCF game and develop a near-optimal pricing strategy for the ECP’s resource pricing problem. Simulation results show that the proposed multi-coalition collaboration mode can improve the system efficiency by $12.64% sim 17.63%$, compared to the traditional single-coalition collaboration mode.
{"title":"An Overlapping Coalition Game Approach for Collaborative Block Mining and Edge Task Offloading in MEC-Assisted Blockchain Networks","authors":"Licheng Ye;Zehui Xiong;Lin Gao;Dusit Niyato","doi":"10.1109/TMC.2025.3591822","DOIUrl":"https://doi.org/10.1109/TMC.2025.3591822","url":null,"abstract":"Mobile edge computing (MEC) is a promising technology that enhances the efficiency of mobile blockchain networks, by enabling miners, often acted by mobile users (MUs) with limited computing resources, to offload resource-intensive mining tasks to nearby edge computing servers. Collaborative block mining can further boost mining efficiency by allowing multiple miners to form <i>coalitions</i>, pooling their computing resources and transaction data together to mine new blocks collaboratively. Therefore, an MEC-assisted collaborative blockchain network can leverage the strengths of both technologies, offering improved efficiency, security, and scalability for blockchain systems. While existing research in this area has mainly focused on the <i>single-coalition</i> collaboration mode, where each miner can only join one coalition, this work explores a more comprehensive <i>multi-coalition</i> collaboration mode, which allows each miner to join multiple coalitions. To analyze the behavior of miners and the edge computing service provider (ECP) in this scenario, we propose a novel two-stage Stackelberg game. In Stage I, the ECP, as the leader, determines the prices of computing resources for all MUs. In Stage II, each MU decides the coalitions to join, resulting in an <i>overlapping coalition formation (OCF) game</i>; Subsequently, each coalition decides how many edge computing resources to purchase from the ECP, leading to an <i>edge resource competition (ERC) game</i>. We derive the closed-form Nash equilibrium for the ERC game, based on which we further propose an OCF-based alternating algorithm to achieve a stable coalition structure for the OCF game and develop a near-optimal pricing strategy for the ECP’s resource pricing problem. Simulation results show that the proposed multi-coalition collaboration mode can improve the system efficiency by <inline-formula><tex-math>$12.64% sim 17.63%$</tex-math></inline-formula>, compared to the traditional single-coalition collaboration mode.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13710-13724"},"PeriodicalIF":9.2,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442731","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-21DOI: 10.1109/TMC.2025.3590799
Zhuangzhuang Zhang;Libing Wu;Zhibo Wang;Jiahui Hu;Chao Ma;Qin Liu
Due to the collaborative machine learning nature of Federated Learning (FL), it enables the training of machine learning models on large-scale distributed datasets in edge computing environments. Nevertheless, the application of FL in edge computing still faces three crucial challenges: resource constraint, privacy leakage, and Byzantine failures. Unfortunately, current approaches lack the ability to effectively balance these three challenges. In this paper, we propose FedEdge, a cost-efficient and secure FL for edge computing. FedEdge contains two main mechanisms: adaptive compression perturbation and dynamic update filtering. The adaptive compression perturbation mechanism reduces the communication overhead, provides different levels of privacy protection for edge nodes, and prevents Byzantine attacks. The dynamic update filtering mechanism is used to further filter Byzantine attacks and limit the impact of adaptive compression perturbation on the global model performance. The experimental results on the MNIST, CIFAR-10, CIFAR-100, and CelebA datasets demonstrate the effectiveness of FedEdge against free-riders, label-flipping, and sign-flipping attacks. Theoretical analysis also demonstrate that FedEdge can still converge even when the majority of edge nodes are malicious.
{"title":"Cost-Efficient and Secure Federated Learning for Edge Computing","authors":"Zhuangzhuang Zhang;Libing Wu;Zhibo Wang;Jiahui Hu;Chao Ma;Qin Liu","doi":"10.1109/TMC.2025.3590799","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590799","url":null,"abstract":"Due to the collaborative machine learning nature of Federated Learning (FL), it enables the training of machine learning models on large-scale distributed datasets in edge computing environments. Nevertheless, the application of FL in edge computing still faces three crucial challenges: resource constraint, privacy leakage, and Byzantine failures. Unfortunately, current approaches lack the ability to effectively balance these three challenges. In this paper, we propose FedEdge, a cost-efficient and secure FL for edge computing. FedEdge contains two main mechanisms: adaptive compression perturbation and dynamic update filtering. The adaptive compression perturbation mechanism reduces the communication overhead, provides different levels of privacy protection for edge nodes, and prevents Byzantine attacks. The dynamic update filtering mechanism is used to further filter Byzantine attacks and limit the impact of adaptive compression perturbation on the global model performance. The experimental results on the MNIST, CIFAR-10, CIFAR-100, and CelebA datasets demonstrate the effectiveness of FedEdge against free-riders, label-flipping, and sign-flipping attacks. Theoretical analysis also demonstrate that FedEdge can still converge even when the majority of edge nodes are malicious.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13615-13632"},"PeriodicalIF":9.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The continuous development of mobile terminal applications, online maps, and other navigation services have become widely used, simultaneously giving rise to significant security risks. To address the issues of privacy leakage and low efficiency in traditional graph shortest path retrieval schemes, an efficient privacy-preserving graph shortest path retrieval scheme is proposed, called EP-GSPR. Specifically, this scheme addresses the privacy security problems in the existing graph shortest path retrieval solutions by ensuring the bilateral privacy protection of the user’s query location and the database privacy of the cloud server. Throughout the retrieval process, the cloud server cannot obtain the user’s location information, and the user cannot access any database information other than the retrieval results. To overcome the performance bottlenecks in existing schemes, a progressive iterative retrieval framework is designed as the fundamental modular, called Pirf, achieving sub-linear retrieval costs and low storage overhead on the cloud server side. Finally, the security analyses demonstrate the EP-GSPR scheme achieves the bilateral privacy-preserving in terms of user and server sides. The comprehensive experiment evaluations also state the efficiency and practicality of the proposed scheme.
{"title":"EP-GSPR: An Efficient Privacy-Preserving Graph Shortest Path Retrieval Scheme","authors":"Chenbin Zhao;Ruifeng Zhu;Jing Chen;Ruiying Du;Kun He;Jianting Ning;Yang Xiang","doi":"10.1109/TMC.2025.3591097","DOIUrl":"https://doi.org/10.1109/TMC.2025.3591097","url":null,"abstract":"The continuous development of mobile terminal applications, online maps, and other navigation services have become widely used, simultaneously giving rise to significant security risks. To address the issues of privacy leakage and low efficiency in traditional graph shortest path retrieval schemes, an efficient privacy-preserving graph shortest path retrieval scheme is proposed, called EP-GSPR. Specifically, this scheme addresses the privacy security problems in the existing graph shortest path retrieval solutions by ensuring the bilateral privacy protection of the user’s query location and the database privacy of the cloud server. Throughout the retrieval process, the cloud server cannot obtain the user’s location information, and the user cannot access any database information other than the retrieval results. To overcome the performance bottlenecks in existing schemes, a progressive iterative retrieval framework is designed as the fundamental modular, called Pirf, achieving sub-linear retrieval costs and low storage overhead on the cloud server side. Finally, the security analyses demonstrate the EP-GSPR scheme achieves the bilateral privacy-preserving in terms of user and server sides. The comprehensive experiment evaluations also state the efficiency and practicality of the proposed scheme.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13633-13647"},"PeriodicalIF":9.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-21DOI: 10.1109/TMC.2025.3590801
Shiyuan Ma;Lei Xie;Chuyu Wang;Yanling Bu;Long Fan;Jingyi Ning;Qing Guo;Baoliu Ye;Sanglu Lu
With the rise of intelligent systems like assisted driving and robotics, all-weather target identification and 3D localization systems have become crucial for reliable obstacle avoidance and navigation. However, vision-based methods struggle to provide accurate target locations under low light or bad weather. Radar-based solutions like mmWave radar and LiDAR are robust but hindered by high costs and challenges in recognizing target identities at scale. In this paper, we propose a low-cost, all-weather target identification and 3D localization system based on LED-tags, which system can address the needs of intelligent systems for obstacle avoidance in complex environments. We explore the backscatter communication of LED devices and design a dual-modal LED-Tag, which includes two features: a backscatter RF signal detectable by RF devices and visual light spot information detectable by cameras, both sharing the same ID. To enhance the limited backscatter capability, we propose a multi-branch parallel model that enhances the signal strength using beamforming synthesis and a channel adjustment mechanism to improve robustness in complex environments, ensuring accurate 3D localization. For multi-target identification, we design an LED-tag encoding system, assigning each tag a unique encoding sequence. Each target’s identity can be recognized with our customized ID decoding method, which leverages prior information and time-domain sampling characteristics. Extensive experimental results show that the backscatter communication and target detection range of LED-tags can reach 15 m. Moreover, the system achieves an average localization error of 7.3 cm within a 5 m range, demonstrating the system’s excellent performance in terms of practicality and accuracy.
{"title":"Multi-Modal Based 3D Localization via the Channel Adjustment LED-Tag","authors":"Shiyuan Ma;Lei Xie;Chuyu Wang;Yanling Bu;Long Fan;Jingyi Ning;Qing Guo;Baoliu Ye;Sanglu Lu","doi":"10.1109/TMC.2025.3590801","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590801","url":null,"abstract":"With the rise of intelligent systems like assisted driving and robotics, all-weather target identification and 3D localization systems have become crucial for reliable obstacle avoidance and navigation. However, vision-based methods struggle to provide accurate target locations under low light or bad weather. Radar-based solutions like mmWave radar and LiDAR are robust but hindered by high costs and challenges in recognizing target identities at scale. In this paper, we propose a <italic>low-cost, all-weather target identification and 3D localization system</i> based on <italic>LED-tags</i>, which system can address the needs of intelligent systems for obstacle avoidance in complex environments. We explore the backscatter communication of LED devices and design a <italic>dual-modal LED-Tag</i>, which includes two features: a backscatter RF signal detectable by RF devices and visual light spot information detectable by cameras, both sharing the same ID. To enhance the limited backscatter capability, we propose a <italic>multi-branch parallel model</i> that enhances the signal strength using beamforming synthesis and a <italic>channel adjustment mechanism</i> to improve robustness in complex environments, ensuring accurate 3D localization. For multi-target identification, we design an LED-tag encoding system, assigning each tag a unique encoding sequence. Each target’s identity can be recognized with our customized <italic>ID decoding method</i>, which leverages prior information and time-domain sampling characteristics. Extensive experimental results show that the backscatter communication and target detection range of LED-tags can reach <italic>15 m</i>. Moreover, the system achieves an <italic>average localization error of 7.3 cm within a 5 m range</i>, demonstrating the system’s excellent performance in terms of practicality and accuracy.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13567-13585"},"PeriodicalIF":9.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-21DOI: 10.1109/TMC.2025.3590969
Jin Yang;Qiong Wu;Zhiying Feng;Zhi Zhou;Deke Guo;Xu Chen
Large Language Models (LLMs) have demonstrated remarkable capabilities, leading to a significant increase in user demand for LLM services. However, cloud-based LLM services often suffer from high latency, unstable responsiveness, and privacy concerns. Therefore, multiple LLMs are usually deployed at the network edge to boost real-time responsiveness and protect data privacy, particularly for many emerging smart mobile and IoT applications. Given the varying response quality and latency of LLM services, a critical issue is how to route user requests from mobile and IoT devices to an appropriate LLM service (i.e., edge LLM expert) to ensure acceptable quality-of-service (QoS). Existing routing algorithms fail to simultaneously address the heterogeneity of LLM services, the interference among requests, and the dynamic workloads necessary for maintaining long-term stable QoS. To meet these challenges, in this paper we propose a novel deep reinforcement learning (DRL)-based QoS-aware LLM routing framework for sustained high-quality LLM services. Due to the dynamic nature of the global state, we propose a dynamic state abstraction technique to compactly represent global state features with a heterogeneous graph attention network (HAN). Additionally, we introduce an action impact estimator and a tailored reward function to guide the DRL agent in maximizing QoS and preventing latency violations. Extensive experiments on both Poisson and real-world workloads demonstrate that our proposed algorithm significantly improves average QoS and computing resource efficiency compared to existing baselines.
{"title":"Quality-of-Service Aware LLM Routing for Edge Computing With Multiple Experts","authors":"Jin Yang;Qiong Wu;Zhiying Feng;Zhi Zhou;Deke Guo;Xu Chen","doi":"10.1109/TMC.2025.3590969","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590969","url":null,"abstract":"Large Language Models (LLMs) have demonstrated remarkable capabilities, leading to a significant increase in user demand for LLM services. However, cloud-based LLM services often suffer from high latency, unstable responsiveness, and privacy concerns. Therefore, multiple LLMs are usually deployed at the network edge to boost real-time responsiveness and protect data privacy, particularly for many emerging smart mobile and IoT applications. Given the varying response quality and latency of LLM services, a critical issue is how to route user requests from mobile and IoT devices to an appropriate LLM service (i.e., edge LLM expert) to ensure acceptable quality-of-service (QoS). Existing routing algorithms fail to simultaneously address the heterogeneity of LLM services, the interference among requests, and the dynamic workloads necessary for maintaining long-term stable QoS. To meet these challenges, in this paper we propose a novel deep reinforcement learning (DRL)-based QoS-aware LLM routing framework for sustained high-quality LLM services. Due to the dynamic nature of the global state, we propose a dynamic state abstraction technique to compactly represent global state features with a heterogeneous graph attention network (HAN). Additionally, we introduce an action impact estimator and a tailored reward function to guide the DRL agent in maximizing QoS and preventing latency violations. Extensive experiments on both Poisson and real-world workloads demonstrate that our proposed algorithm significantly improves average QoS and computing resource efficiency compared to existing baselines.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13648-13662"},"PeriodicalIF":9.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Smart grid, through networked smart meters employing the non-intrusive load monitoring (NILM) technique, can considerably discern the usage patterns of residential appliances. However, this technique also incurs privacy leakage. To address this issue, we propose an innovative scheme based on adversarial attack in this paper. The scheme effectively prevents NILM models from violating appliance-level privacy, while also ensuring accurate billing calculation for users. To achieve this objective, we overcome two primary challenges. First, as NILM models fall under the category of time-series regression models, direct application of traditional adversarial attacks designed for classification tasks is not feasible. To tackle this issue, we formulate a novel adversarial attack problem tailored specifically for NILM and providing a theoretical foundation for utilizing the Jacobian of the NILM model to generate imperceptible perturbations. Leveraging the Jacobian, our scheme can produce perturbations, which effectively misleads the signal prediction of NILM models to safeguard users’ appliance-level privacy. The second challenge pertains to fundamental utility requirements, where existing adversarial attack schemes struggle to achieve accurate billing calculation for users. To handle this problem, we introduce an additional constraint, mandating that the sum of added perturbations within a billing period must be precisely zero. Experimental validation on real-world power datasets REDD and U.K.-DALE demonstrates the efficacy of our proposed solutions, which can significantly amplify the discrepancy between the output of the targeted NILM model and the actual power signal of appliances, and enable accurate billing at the same time. Additionally, our solutions exhibit transferability, making the generated perturbation signal from one target model applicable to other diverse NILM models.
{"title":"Preventing Non-Intrusive Load Monitoring Privacy Invasion: A Precise Adversarial Attack Scheme for Networked Smart Meters","authors":"Jialing He;Jiacheng Wang;Ning Wang;Shangwei Guo;Liehuang Zhu;Dusit Niyato;Tao Xiang","doi":"10.1109/TMC.2025.3590765","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590765","url":null,"abstract":"Smart grid, through networked smart meters employing the non-intrusive load monitoring (NILM) technique, can considerably discern the usage patterns of residential appliances. However, this technique also incurs privacy leakage. To address this issue, we propose an innovative scheme based on adversarial attack in this paper. The scheme effectively prevents NILM models from violating appliance-level privacy, while also ensuring accurate billing calculation for users. To achieve this objective, we overcome two primary challenges. First, as NILM models fall under the category of time-series regression models, direct application of traditional adversarial attacks designed for classification tasks is not feasible. To tackle this issue, we formulate a novel adversarial attack problem tailored specifically for NILM and providing a theoretical foundation for utilizing the Jacobian of the NILM model to generate imperceptible perturbations. Leveraging the Jacobian, our scheme can produce perturbations, which effectively misleads the signal prediction of NILM models to safeguard users’ appliance-level privacy. The second challenge pertains to fundamental utility requirements, where existing adversarial attack schemes struggle to achieve accurate billing calculation for users. To handle this problem, we introduce an additional constraint, mandating that the sum of added perturbations within a billing period must be precisely zero. Experimental validation on real-world power datasets REDD and U.K.-DALE demonstrates the efficacy of our proposed solutions, which can significantly amplify the discrepancy between the output of the targeted NILM model and the actual power signal of appliances, and enable accurate billing at the same time. Additionally, our solutions exhibit transferability, making the generated perturbation signal from one target model applicable to other diverse NILM models.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13696-13709"},"PeriodicalIF":9.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442691","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-18DOI: 10.1109/TMC.2025.3590606
Hongjun Wang;Jiyuan Chen;Tong Pan;Zheng Dong;Renhe Jiang;Xuan Song
Spatiotemporal neural networks have shown great promise in urban scenarios by effectively capturing temporal and spatial correlations. However, urban environments are constantly evolving, and current model evaluations are often limited to traffic scenarios and use data mainly collected only a few weeks after training period to evaluate model performance. The generalization ability of these models remains largely unexplored. To address this, we propose a Spatiotemporal Out-of-Distribution (ST-OOD) benchmark, which comprises six urban scenario: bike-sharing, 311 services, pedestrian counts, traffic speed, traffic flow, ride-hailing demand, and bike-sharing, each with in-distribution (same year) and out-of-distribution (next years) settings. We extensively evaluate state-of-the-art spatiotemporal models and find that their performance degrades significantly in out-of-distribution settings, with most models performing even worse than a simple Multi-Layer Perceptron (MLP). Our findings suggest that current leading methods tend to over-rely on parameters to overfit training data, which may lead to good performance on in-distribution data but often results in poor generalization. We also investigated whether dropout could mitigate the negative effects of overfitting. Our results showed that a slight dropout rate could significantly improve generalization performance on most datasets, with minimal impact on in-distribution performance. However, balancing in-distribution and out-of-distribution performance remains a challenging problem. We hope that the proposed benchmark will encourage further research on this critical issue.
{"title":"Evaluating the Generalization Ability of Spatiotemporal Model in Urban Scenario","authors":"Hongjun Wang;Jiyuan Chen;Tong Pan;Zheng Dong;Renhe Jiang;Xuan Song","doi":"10.1109/TMC.2025.3590606","DOIUrl":"https://doi.org/10.1109/TMC.2025.3590606","url":null,"abstract":"Spatiotemporal neural networks have shown great promise in urban scenarios by effectively capturing temporal and spatial correlations. However, urban environments are constantly evolving, and current model evaluations are often limited to traffic scenarios and use data mainly collected only a few weeks after training period to evaluate model performance. The generalization ability of these models remains largely unexplored. To address this, we propose a Spatiotemporal Out-of-Distribution (ST-OOD) benchmark, which comprises six urban scenario: bike-sharing, 311 services, pedestrian counts, traffic speed, traffic flow, ride-hailing demand, and bike-sharing, each with in-distribution (same year) and out-of-distribution (next years) settings. We extensively evaluate state-of-the-art spatiotemporal models and find that their performance degrades significantly in out-of-distribution settings, with most models performing even worse than a simple Multi-Layer Perceptron (MLP). Our findings suggest that current leading methods tend to over-rely on parameters to overfit training data, which may lead to good performance on in-distribution data but often results in poor generalization. We also investigated whether dropout could mitigate the negative effects of overfitting. Our results showed that a slight dropout rate could significantly improve generalization performance on most datasets, with minimal impact on in-distribution performance. However, balancing in-distribution and out-of-distribution performance remains a challenging problem. We hope that the proposed benchmark will encourage further research on this critical issue.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 12","pages":"13535-13548"},"PeriodicalIF":9.2,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145442692","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-10DOI: 10.1109/TMC.2025.3586793
Hebin Huang;Junbin Liang;Geyong Min
Edge intelligence is an emerging paradigm in edge computing that deploys Deep Neural Network (DNN) models on edge servers with limited storage and computation capacities to provide inference services for high mobility and real-time applications, such as autonomous driving or smart surveillance, with varying accuracy and delay requirements. Adapting application configurations (e.g., image resolution or video frame rate) while selecting different DNN models and deployment locations can provide high-accuracy, low-delay inference services that meet user requirements. However, the configurations and DNN models of various inference services are highly heterogeneous. As balancing inference accuracy, resource cost, and delay is a multi-objective programming problem, it is a great challenge to obtain the optimal solution. To address this challenge, we propose a novel online framework to jointly optimize the configuration adaption, DNN model selection, and deployment for heterogeneous inference services. Specifically, we first formulate this joint optimization problem as an integer linear programming problem and prove it is NP-hard. Then, we further model the problem as a Partial Observable Markov Decision Process (POMDP) and solve it by developing a Heterogeneous-Agent Reinforcement Learning (HARL) based algorithm, named Heterogeneous Inference Service ProvidER (HISPER). It allows agents to have different action spaces corresponding to different types of configurations and DNN models. Finally, extensive experiments demonstrate that the proposed algorithm outperforms other state-of-the-art counterparts.
{"title":"Joint DNN Model Deployment, Selection, and Configuration for Heterogeneous Inference Services Toward Edge Intelligence","authors":"Hebin Huang;Junbin Liang;Geyong Min","doi":"10.1109/TMC.2025.3586793","DOIUrl":"https://doi.org/10.1109/TMC.2025.3586793","url":null,"abstract":"Edge intelligence is an emerging paradigm in edge computing that deploys Deep Neural Network (DNN) models on edge servers with limited storage and computation capacities to provide inference services for high mobility and real-time applications, such as autonomous driving or smart surveillance, with varying accuracy and delay requirements. Adapting application configurations (e.g., image resolution or video frame rate) while selecting different DNN models and deployment locations can provide high-accuracy, low-delay inference services that meet user requirements. However, the configurations and DNN models of various inference services are highly heterogeneous. As balancing inference accuracy, resource cost, and delay is a multi-objective programming problem, it is a great challenge to obtain the optimal solution. To address this challenge, we propose a novel online framework to jointly optimize the configuration adaption, DNN model selection, and deployment for heterogeneous inference services. Specifically, we first formulate this joint optimization problem as an integer linear programming problem and prove it is NP-hard. Then, we further model the problem as a Partial Observable Markov Decision Process (POMDP) and solve it by developing a Heterogeneous-Agent Reinforcement Learning (HARL) based algorithm, named Heterogeneous Inference Service ProvidER (HISPER). It allows agents to have different action spaces corresponding to different types of configurations and DNN models. Finally, extensive experiments demonstrate that the proposed algorithm outperforms other state-of-the-art counterparts.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 11","pages":"12726-12741"},"PeriodicalIF":9.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145223671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-10DOI: 10.1109/TMC.2025.3586626
Hefan Zhang;Zhiyuan Wang;Wenhao Lu;Shan Zhang;Hongbin Luo
Low-earth-orbit (LEO) mega-constellations with inter-satellite links (ISLs) are becoming the Internet backbone in space. Satellites within LEO often need the capability to enforce data forwarding paths. For example, they may need to bypass the satellites over the untrusted areas for the data of mission-critical applications or minimize latency for the data of time-sensitive applications. However, typical source/segment routing techniques (e.g., SRv6) suffer from scalability issue, since they record source-route-style forwarding information via the list-based structure. This results in great payload and forwarding overhead. To overcome this drawback, we propose a source/segment routing architecture for LEO mega-constellations, which is named as Link-identified Routing (LiR). LiR leverages in-packet bloom filter (BF) to record source-route-style forwarding information. BF could efficiently record multiple elements via a probabilistic data structure, but overlooks the order of the encoded elements. To address this, LiR identifies each unidirectional ISL, and represents the path by encoding ISL identifiers into BF. We investigate how to optimize BF configuration and ISL encoding policy to address false positives caused by BF. We implement LiR in Linux kernel and develop a container-based emulator for performance evaluation. Results show that LiR significantly outperforms SRv6 in terms of packet forwarding and data delivery efficiency.
{"title":"Source Routing for LEO Mega-Constellations Based on Bloom Filter","authors":"Hefan Zhang;Zhiyuan Wang;Wenhao Lu;Shan Zhang;Hongbin Luo","doi":"10.1109/TMC.2025.3586626","DOIUrl":"https://doi.org/10.1109/TMC.2025.3586626","url":null,"abstract":"Low-earth-orbit (LEO) mega-constellations with inter-satellite links (ISLs) are becoming the Internet backbone in space. Satellites within LEO often need the capability to enforce data forwarding paths. For example, they may need to bypass the satellites over the untrusted areas for the data of mission-critical applications or minimize latency for the data of time-sensitive applications. However, typical source/segment routing techniques (e.g., SRv6) suffer from scalability issue, since they record source-route-style forwarding information via the list-based structure. This results in great payload and forwarding overhead. To overcome this drawback, we propose a source/segment routing architecture for LEO mega-constellations, which is named as Link-identified Routing (LiR). LiR leverages in-packet bloom filter (BF) to record source-route-style forwarding information. BF could efficiently record multiple elements via a probabilistic data structure, but overlooks the order of the encoded elements. To address this, LiR identifies each unidirectional ISL, and represents the path by encoding ISL identifiers into BF. We investigate how to optimize BF configuration and ISL encoding policy to address false positives caused by BF. We implement LiR in Linux kernel and develop a container-based emulator for performance evaluation. Results show that LiR significantly outperforms SRv6 in terms of packet forwarding and data delivery efficiency.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 11","pages":"12487-12504"},"PeriodicalIF":9.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145223679","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-07DOI: 10.1109/TMC.2025.3586644
Jun Liu;Yunming Liao;Hongli Xu;Yang Xu;Jianchun Liu;Chen Qian
Federated fine-tuning (FedFT) has been proposed to fine-tune the pre-trained language models in a distributed manner. However, there are two critical challenges for efficient FedFT in practical applications, i.e., resource constraints and system heterogeneity. Existing works rely on parameter-efficient fine-tuning methods, e.g., low-rank adaptation (LoRA)1, but with major limitations. Herein, based on the inherent characteristics of FedFT, we observe that LoRA layers with higher ranks added close to the output help to save resource consumption while achieving comparable fine-tuning performance. Then we propose a novel LoRA-based FedFT framework, termed LEGEND, which faces the difficulty of determining the number of LoRA layers (called, LoRA depth) and the rank of each LoRA layer (called, rank distribution). We analyze the coupled relationship between LoRA depth and rank distribution, and design an efficient LoRA configuration algorithm for heterogeneous devices, thereby promoting fine-tuning efficiency. Extensive experiments are conducted on a physical platform with 80 commercial devices. The results show that LEGEND can achieve a speedup of 1.5-2.8× and save communication costs by about 42.3% when achieving the target accuracy, compared to the advanced solutions.
{"title":"Adaptive Parameter-Efficient Federated Fine-Tuning on Heterogeneous Devices","authors":"Jun Liu;Yunming Liao;Hongli Xu;Yang Xu;Jianchun Liu;Chen Qian","doi":"10.1109/TMC.2025.3586644","DOIUrl":"https://doi.org/10.1109/TMC.2025.3586644","url":null,"abstract":"Federated fine-tuning (FedFT) has been proposed to fine-tune the pre-trained language models in a distributed manner. However, there are two critical challenges for efficient FedFT in practical applications, i.e., resource constraints and system heterogeneity. Existing works rely on parameter-efficient fine-tuning methods, e.g., low-rank adaptation (LoRA)<sup>1</sup>, but with major limitations. Herein, based on the inherent characteristics of FedFT, we observe that LoRA layers with higher ranks added close to the output help to save resource consumption while achieving comparable fine-tuning performance. Then we propose a novel LoRA-based FedFT framework, termed LEGEND, which faces the difficulty of determining the number of LoRA layers (called, LoRA depth) and the rank of each LoRA layer (called, rank distribution). We analyze the coupled relationship between LoRA depth and rank distribution, and design an efficient LoRA configuration algorithm for heterogeneous devices, thereby promoting fine-tuning efficiency. Extensive experiments are conducted on a physical platform with 80 commercial devices. The results show that LEGEND can achieve a speedup of 1.5-2.8× and save communication costs by about 42.3% when achieving the target accuracy, compared to the advanced solutions.","PeriodicalId":50389,"journal":{"name":"IEEE Transactions on Mobile Computing","volume":"24 11","pages":"12533-12549"},"PeriodicalIF":9.2,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221219","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","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：