Yanghe Pan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yanghe Pan has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Yuntao Wang
Xi'an Jiaotong University, Xi'an, China
16 shared publications - Zhou Su
Xi’an Jiaotong University
16 shared publications - Ruidong Li
Institute of Science and Engineering, Kanazawa University, Kakuma, Kanazawa, Japan
6 shared publications - Shaolong Guo
School of Cyber Science and Engineering, Xian Jiaotong University, Xian, China
6 shared publications - Tom Hao Luan
School of Cyber Science and Engineering, Xian Jiaotong University, Xian, China
6 shared publications - Jianbing Ni
Queen’s University
3 shared publications
Latest Publications
- 10 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Network
- Vol. PP (08908044) , 1-9
Agent Discovery in Internet of Agents: Challenges and Solutions
Recent advances in large language models and agentic AI are enabling the emergence of large-scale agent ecosystems, commonly referred to as the Internet of Agents (IoA), where diverse software and embodied agents interact and collaborate to accomplish complex tasks. A key prerequisite for such large-scale collaboration is agent capability discovery, where agents identify, advertise, and match one another’s capabilities under dynamic tasks. Agent’s capability in IoA is inherently heterogeneous and context-dependent, raising challenges in capability representation, scalable discovery, and long-term performance. To address these issues, this paper introduces a novel two-stage capability discovery framework. The first stage, autonomous capability announcement, allows agents to credibly publish machine-interpretable descriptions of their abilities. The second stage, task-driven capability discovery, enables context-aware search, ranking, and composition to locate and assemble suitable agents for specific tasks. Building on this framework, we propose a novel scheme that integrates semantic capability modeling, scalable and updatable indexing, and memory-enhanced continual discovery. Simulation results demonstrate that our approach enhances discovery performance and scalability. Finally, we outline a research roadmap and highlight open problems and promising directions for future IoA.
- 10 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Network
- Vol. PP (08908044) , 1-9
Agent Discovery in Internet of Agents: Challenges and Solutions
Recent advances in large language models and agentic AI are enabling the emergence of large-scale agent ecosystems, commonly referred to as the Internet of Agents (IoA), where diverse software and embodied agents interact and collaborate to accomplish complex tasks. A key prerequisite for such large-scale collaboration is agent capability discovery, where agents identify, advertise, and match one another’s capabilities under dynamic tasks. Agent’s capability in IoA is inherently heterogeneous and context-dependent, raising challenges in capability representation, scalable discovery, and long-term performance. To address these issues, this paper introduces a novel two-stage capability discovery framework. The first stage, autonomous capability announcement, allows agents to credibly publish machine-interpretable descriptions of their abilities. The second stage, task-driven capability discovery, enables context-aware search, ranking, and composition to locate and assemble suitable agents for specific tasks. Building on this framework, we propose a novel scheme that integrates semantic capability modeling, scalable and updatable indexing, and memory-enhanced continual discovery. Simulation results demonstrate that our approach enhances discovery performance and scalability. Finally, we outline a research roadmap and highlight open problems and promising directions for future IoA.
- 27 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3995-4010
FLET: Game-Theoretic Free-Riding Mitigation via Test Tasks in Federated Learning
Federated learning (FL) is a widely studied framework for privacy-preserving collaborative training among multiple clients. However, real-world deployments reveal a persistent challenge: free-riders, i.e., participants who benefit from the system without contributing meaningful updates. If not properly addressed, free-riders can discourage honest contributors and ultimately impair both the fairness and efficiency of FL ecosystems. Existing defenses mainly rely on post-hoc per-client, per-round evaluation, leading to limited deterrence and high resource overhead, particularly in large-scale deployments. To tackle these problems, we propose FLET, a novel federated learning framework with test tasks. FLET introduces dedicated test tasks into the training process, blending them with real FL tasks while concealing their types from participants. These test tasks, generated from the reference datasets, serve as decoys that enable accurate detection of free-riding behavior. To enforce accountability, an economic penalty mechanism is employed, achieving both proactive (ex-ante) deterrence and reactive (expost) detection.We analyze the interactions between the server and participants through a free-riding suppression game model with asymmetric information (i.e., task type), and develop a strategic information disclosure scheme (i.e., revealing task requirements) to mislead attackers and proactively shape participant behavior. We characterize both pure-strategy and mixed-strategy perfect Bayesian Nash equilibria, and propose a lightweight strateg-ymaking algorithm that guides players toward equilibrium strategies under different conditions with modest overhead. Extensive experiments validate that FLET effectively suppresses free-riding and enhances the utility of both the server and participants. Our findings provide insights for designing cost-effective free-riding defenses in practical FL.
- 27 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3995-4010
FLET: Game-Theoretic Free-Riding Mitigation via Test Tasks in Federated Learning
Federated learning (FL) is a widely studied framework for privacy-preserving collaborative training among multiple clients. However, real-world deployments reveal a persistent challenge: free-riders, i.e., participants who benefit from the system without contributing meaningful updates. If not properly addressed, free-riders can discourage honest contributors and ultimately impair both the fairness and efficiency of FL ecosystems. Existing defenses mainly rely on post-hoc per-client, per-round evaluation, leading to limited deterrence and high resource overhead, particularly in large-scale deployments. To tackle these problems, we propose FLET, a novel federated learning framework with test tasks. FLET introduces dedicated test tasks into the training process, blending them with real FL tasks while concealing their types from participants. These test tasks, generated from the reference datasets, serve as decoys that enable accurate detection of free-riding behavior. To enforce accountability, an economic penalty mechanism is employed, achieving both proactive (ex-ante) deterrence and reactive (expost) detection.We analyze the interactions between the server and participants through a free-riding suppression game model with asymmetric information (i.e., task type), and develop a strategic information disclosure scheme (i.e., revealing task requirements) to mislead attackers and proactively shape participant behavior. We characterize both pure-strategy and mixed-strategy perfect Bayesian Nash equilibria, and propose a lightweight strateg-ymaking algorithm that guides players toward equilibrium strategies under different conditions with modest overhead. Extensive experiments validate that FLET effectively suppresses free-riding and enhances the utility of both the server and participants. Our findings provide insights for designing cost-effective free-riding defenses in practical FL.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3242-3257
Privacy-Utility Trade-Off in Federated LLM Fine-Tuning: A Dynamic Game Approach
Fine-tuning large language models (LLMs) is critical for adapting pretrained models to specialized downstream tasks. Federated LLM fine-tuning enables privacy-aware model updates by allowing data owners (DOs) to contribute a global LLM without exposing local data. However, full-parameter fine-tuning in federated settings incurs significant computational and communication overhead, while frequent gradient exchanges increase the risk of privacy leakage, such as memorized data inference. Parameter-efficient fine-tuning (PEFT) with differential privacy (DP) offers a low-overhead alternative with formal privacy guarantees, but fails to strike privacy-utility tradeoff under heterogeneous privacy preferences: individual DOs may inject excessive DP noise to maximize privacy, whereas the curator aims to minimize noise to preserve model quality. In this paper, we present an innovative game-theoretical framework that enables dynamic privacy trading within differentially private federated LLM fine-tuning. In the game, DOs strategically adjust their local DP noise levels in exchange for customized incentives from the curator, thereby balancing privacy and utility. We begin by establishing a theoretical convergence bound that quantifies the influence of locally injected noise on the global model utility. Under this bound, we analytically characterize the pure-strategy Nash equilibrium of the game, accounting for DO heterogeneity, curator budget constraints, and noise estimation errors. For mixed-strategy settings with incomplete information, we design a hierarchical reinforcement learning algorithm that jointly learns DOs’ optimal noise-saving strategies and the curator’s optimal pricing policy without presupposing their private information. Experiments on real-world datasets demonstrate that the proposed scheme improves DO utility, reduces curator cost, mitigates free-riding, and accelerates convergence compared to existing methods.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3242-3257
Privacy-Utility Trade-Off in Federated LLM Fine-Tuning: A Dynamic Game Approach
Fine-tuning large language models (LLMs) is critical for adapting pretrained models to specialized downstream tasks. Federated LLM fine-tuning enables privacy-aware model updates by allowing data owners (DOs) to contribute a global LLM without exposing local data. However, full-parameter fine-tuning in federated settings incurs significant computational and communication overhead, while frequent gradient exchanges increase the risk of privacy leakage, such as memorized data inference. Parameter-efficient fine-tuning (PEFT) with differential privacy (DP) offers a low-overhead alternative with formal privacy guarantees, but fails to strike privacy-utility tradeoff under heterogeneous privacy preferences: individual DOs may inject excessive DP noise to maximize privacy, whereas the curator aims to minimize noise to preserve model quality. In this paper, we present an innovative game-theoretical framework that enables dynamic privacy trading within differentially private federated LLM fine-tuning. In the game, DOs strategically adjust their local DP noise levels in exchange for customized incentives from the curator, thereby balancing privacy and utility. We begin by establishing a theoretical convergence bound that quantifies the influence of locally injected noise on the global model utility. Under this bound, we analytically characterize the pure-strategy Nash equilibrium of the game, accounting for DO heterogeneity, curator budget constraints, and noise estimation errors. For mixed-strategy settings with incomplete information, we design a hierarchical reinforcement learning algorithm that jointly learns DOs’ optimal noise-saving strategies and the curator’s optimal pricing policy without presupposing their private information. Experiments on real-world datasets demonstrate that the proposed scheme improves DO utility, reduces curator cost, mitigates free-riding, and accelerates convergence compared to existing methods.
- 10 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Vehicular Technology Magazine
- Vol. 21 (2) , 110-119
Trustworthy Semantic Communication for Vehicular Networks: Challenges and Solutions
Semantic communication (SemCom) has the potential to significantly reduce communication delay in vehicle-to-everything (V2X) communications within vehicular networks (VNs). However, the deployment of vehicular SemCom networks (VN-SemComNets) faces critical trust challenges in information transmission, semantic encoding, and communication entity reliability. This article proposes an innovative three-layer trustworthy VN-SemComNet architecture. Specifically, we introduce a semantic camouflage transmission mechanism leveraging defensive adversarial noise for active eavesdropping defense, a robust federated encoder-decoder training framework to mitigate encoder-decoder poisoning attacks, and an audit game-based distributed vehicle trust management mechanism to deter untrustworthy vehicles. A case study validates the effectiveness of the proposed solutions. Lastly, essential future research directions are pointed out to advance this emerging field.
- 10 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Vehicular Technology Magazine
- Vol. 21 (2) , 110-119
Trustworthy Semantic Communication for Vehicular Networks: Challenges and Solutions
Semantic communication (SemCom) has the potential to significantly reduce communication delay in vehicle-to-everything (V2X) communications within vehicular networks (VNs). However, the deployment of vehicular SemCom networks (VN-SemComNets) faces critical trust challenges in information transmission, semantic encoding, and communication entity reliability. This article proposes an innovative three-layer trustworthy VN-SemComNet architecture. Specifically, we introduce a semantic camouflage transmission mechanism leveraging defensive adversarial noise for active eavesdropping defense, a robust federated encoder-decoder training framework to mitigate encoder-decoder poisoning attacks, and an audit game-based distributed vehicle trust management mechanism to deter untrustworthy vehicles. A case study validates the effectiveness of the proposed solutions. Lastly, essential future research directions are pointed out to advance this emerging field.
- Top Cited Papers
- 20 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 4476-4501
Internet of Agents: Fundamentals, Applications, and Challenges
With the rapid proliferation of large language models and vision‑language models, AI agents have evolved from isolated, task‑specific systems into autonomous, interactive entities capable of perceiving, reasoning, and acting without human intervention. As these agents proliferate across virtual and physical environments, from virtual assistants to embodied robots, the need for a unified, agent‑centric infrastructure becomes paramount. In this survey, we introduce the Internet of Agents (IoA) as a foundational framework that enables seamless interconnection, dynamic discovery, and collaborative orchestration among heterogeneous agents at scale. We begin by presenting a general IoA architecture, highlighting its hierarchical organization, distinguishing features relative to the traditional Internet, and emerging applications. Next, we analyze the key operational enablers of IoA, including capability notification and discovery, adaptive communication protocols, dynamic task matching, consensus and conflict‑resolution mechanisms, and incentive models. Finally, we identify open research directions toward building resilient and trustworthy IoA ecosystems.
- 20 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 4476-4501
Top Cited PapersInternet of Agents: Fundamentals, Applications, and Challenges
With the rapid proliferation of large language models and vision‑language models, AI agents have evolved from isolated, task‑specific systems into autonomous, interactive entities capable of perceiving, reasoning, and acting without human intervention. As these agents proliferate across virtual and physical environments, from virtual assistants to embodied robots, the need for a unified, agent‑centric infrastructure becomes paramount. In this survey, we introduce the Internet of Agents (IoA) as a foundational framework that enables seamless interconnection, dynamic discovery, and collaborative orchestration among heterogeneous agents at scale. We begin by presenting a general IoA architecture, highlighting its hierarchical organization, distinguishing features relative to the traditional Internet, and emerging applications. Next, we analyze the key operational enablers of IoA, including capability notification and discovery, adaptive communication protocols, dynamic task matching, consensus and conflict‑resolution mechanisms, and incentive models. Finally, we identify open research directions toward building resilient and trustworthy IoA ecosystems.
- 11 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1616-1631
The Right to Be Forgotten Versus the Need to Be Remembered: Efficient Personalized Federated Unlearning With Optimal Incentives
Federated Learning (FL) enables privacy-preserving collaborative training across distributed devices without direct data sharing. However, existing FL approaches face new challenges in upholding users' right to be forgotten (RTBF), particularly in reconciling data removal guarantees with model performance preservation. Current server-centric federated unlearning (FU) approaches suffer from two critical limitations: i) inability to accommodate personalized unlearning requirements across users, and ii) lack of verifiable data erasure guarantees. We propose a novel user-centric personalized FU (PFU) framework that shifts unlearning control to data owners while maintaining global model utility, where they independently and adaptively adjust unlearning intensity parameters (e.g., distillation temperature) to meet personalized forgetting requirements through multi-task alternating training on local data and unlearning knowledge distillation. To further resolve the fundamental conflict between users' RTBF and servers' need to be remembered (NTBR), a contract-theoretic incentive mechanism is devised to compensate users for reducing their unlearning intensity, thereby enhancing model utility. We theoretically prove the optimality of the designed contracts under multi-dimensional information asymmetry, which maximize the server's utility and motivate users to truthfully select and sign the corresponding contract items that align with their privacy expectations. Extensive experiments demonstrate that our proposed scheme significantly enhances the performance of unlearned global models and improves the utilities of both the server and the users compared to existing approaches.
- 11 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1616-1631
The Right to Be Forgotten Versus the Need to Be Remembered: Efficient Personalized Federated Unlearning With Optimal Incentives
Federated Learning (FL) enables privacy-preserving collaborative training across distributed devices without direct data sharing. However, existing FL approaches face new challenges in upholding users' right to be forgotten (RTBF), particularly in reconciling data removal guarantees with model performance preservation. Current server-centric federated unlearning (FU) approaches suffer from two critical limitations: i) inability to accommodate personalized unlearning requirements across users, and ii) lack of verifiable data erasure guarantees. We propose a novel user-centric personalized FU (PFU) framework that shifts unlearning control to data owners while maintaining global model utility, where they independently and adaptively adjust unlearning intensity parameters (e.g., distillation temperature) to meet personalized forgetting requirements through multi-task alternating training on local data and unlearning knowledge distillation. To further resolve the fundamental conflict between users' RTBF and servers' need to be remembered (NTBR), a contract-theoretic incentive mechanism is devised to compensate users for reducing their unlearning intensity, thereby enhancing model utility. We theoretically prove the optimality of the designed contracts under multi-dimensional information asymmetry, which maximize the server's utility and motivate users to truthfully select and sign the corresponding contract items that align with their privacy expectations. Extensive experiments demonstrate that our proposed scheme significantly enhances the performance of unlearned global models and improves the utilities of both the server and the users compared to existing approaches.
- Top Cited PapersOpen Access
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Computer Society
- Vol. 6, 1611-1624
Security of Internet of Agents: Attacks and Countermeasures
With the rise of large language and vision-language models, AI agents have evolved into autonomous, interactive systems capable of perception, reasoning, and decision-making. As they proliferate across virtual and physical domains, the Internet of Agents (IoA) has emerged as a key infrastructure for enabling scalable and secure coordination among heterogeneous agents. This survey offers a comprehensive examination of the security and privacy landscape in IoA systems. We begin by outlining the IoA architecture and its distinct vulnerabilities compared to traditional networks, focusing on four critical aspects: identity authentication threats, cross-agent trust issues, embodied security, and privacy risks. We then review existing and emerging defense mechanisms and highlight persistent challenges. Finally, we identify open research directions to advance the development of resilient and privacy-preserving IoA ecosystems.
- 16 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Computer Society
- Vol. 6, 1611-1624
Top Cited PapersOpen AccessSecurity of Internet of Agents: Attacks and Countermeasures
With the rise of large language and vision-language models, AI agents have evolved into autonomous, interactive systems capable of perception, reasoning, and decision-making. As they proliferate across virtual and physical domains, the Internet of Agents (IoA) has emerged as a key infrastructure for enabling scalable and secure coordination among heterogeneous agents. This survey offers a comprehensive examination of the security and privacy landscape in IoA systems. We begin by outlining the IoA architecture and its distinct vulnerabilities compared to traditional networks, focusing on four critical aspects: identity authentication threats, cross-agent trust issues, embodied security, and privacy risks. We then review existing and emerging defense mechanisms and highlight persistent challenges. Finally, we identify open research directions to advance the development of resilient and privacy-preserving IoA ecosystems.
- Top Cited Papers
- 3 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 28, 1906-1949
Large Model-Based Agents: State-of-the-Art, Cooperation Paradigms, Security and Privacy, and Future Trends
With the rapid advancement of large models (LMs), the development of general-purpose intelligent agents powered by LMs has become a reality. It is foreseeable that in the near future, LM-driven general AI agents will serve as essential tools in production tasks, capable of autonomous communication and collaboration without human intervention. This paper investigates scenarios involving the autonomous collaboration of future LM agents. We review the current state of LM agents, the key technologies enabling LM agent collaboration, and the security and privacy challenges they face during cooperative operations. To this end, we first explore the foundational principles of LM agents, including their general architecture, key components, enabling technologies, and modern applications. We then discuss practical collaboration paradigms from data, computation, and knowledge perspectives to achieve connected intelligence among LM agents. After that, we analyze the security vulnerabilities and privacy risks associated with LM agents, particularly in multi-agent settings, examining underlying mechanisms and reviewing current and potential countermeasures. Lastly, we propose future research directions for building robust and secure LM agent ecosystems.
- 3 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 28, 1906-1949
Top Cited PapersLarge Model-Based Agents: State-of-the-Art, Cooperation Paradigms, Security and Privacy, and Future Trends
With the rapid advancement of large models (LMs), the development of general-purpose intelligent agents powered by LMs has become a reality. It is foreseeable that in the near future, LM-driven general AI agents will serve as essential tools in production tasks, capable of autonomous communication and collaboration without human intervention. This paper investigates scenarios involving the autonomous collaboration of future LM agents. We review the current state of LM agents, the key technologies enabling LM agent collaboration, and the security and privacy challenges they face during cooperative operations. To this end, we first explore the foundational principles of LM agents, including their general architecture, key components, enabling technologies, and modern applications. We then discuss practical collaboration paradigms from data, computation, and knowledge perspectives to achieve connected intelligence among LM agents. After that, we analyze the security vulnerabilities and privacy risks associated with LM agents, particularly in multi-agent settings, examining underlying mechanisms and reviewing current and potential countermeasures. Lastly, we propose future research directions for building robust and secure LM agent ecosystems.
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (6) , 9461-9475
Privacy-Preserving Byzantine-Robust Federated Learning via Deep Reinforcement Learning in Vehicular Networks
Federated learning (FL), as a transformative approach in vehicular networks, enables collaborative model training without exposure of vehicle users' local data. However, vehicular FL services are susceptible to Byzantine attacks, since malicious vehicles can upload false local updates to degrade the performance of the global model. Additionally, the honest-but-curious mobile edge computing (MEC) nodes may attempt to extract sensitive information from the shared updates via inference or inversion attacks. In this paper, we propose DRL-PBFL, a privacy-preserving Byzantine-robust FL scheme in vehicular networks that tolerates Byzantine vehicles and the honest-but-curious MEC node during the training process. Specifically, the deep reinforcement learning (DRL) technique is leveraged to maintain the robustness of FL against Byzantine vehicles, and a novel secure aggregation algorithm is designed to prevent curious inferences from the honest-but-curious MEC node. A performance-based weighting aggregation policy optimized by a deep deterministic policy gradients (DDPG) component is also devised to aggregate local updates. As it is challenging to directly integrate weighting aggregation policy with general pseudo-random masking methods for Byzantine resistance, the Lagrange interpolation is utilized to generate local perturbations to defend against the honest-but-curious MEC node and eliminate the weighted perturbations after secure aggregation. Notably, DRL-PBFL effectively achieves Byzantine robustness in FL even though the vehicles' data is not independently and identically distributed (non-IID). Experiments on three benchmark datasets with different adversarial settings show that DRL-PBFL effectively improves the global model accuracy under tailored Byzantine attacks compared with state-of-the-art schemes.
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (6) , 9461-9475
Privacy-Preserving Byzantine-Robust Federated Learning via Deep Reinforcement Learning in Vehicular Networks
Federated learning (FL), as a transformative approach in vehicular networks, enables collaborative model training without exposure of vehicle users' local data. However, vehicular FL services are susceptible to Byzantine attacks, since malicious vehicles can upload false local updates to degrade the performance of the global model. Additionally, the honest-but-curious mobile edge computing (MEC) nodes may attempt to extract sensitive information from the shared updates via inference or inversion attacks. In this paper, we propose DRL-PBFL, a privacy-preserving Byzantine-robust FL scheme in vehicular networks that tolerates Byzantine vehicles and the honest-but-curious MEC node during the training process. Specifically, the deep reinforcement learning (DRL) technique is leveraged to maintain the robustness of FL against Byzantine vehicles, and a novel secure aggregation algorithm is designed to prevent curious inferences from the honest-but-curious MEC node. A performance-based weighting aggregation policy optimized by a deep deterministic policy gradients (DDPG) component is also devised to aggregate local updates. As it is challenging to directly integrate weighting aggregation policy with general pseudo-random masking methods for Byzantine resistance, the Lagrange interpolation is utilized to generate local perturbations to defend against the honest-but-curious MEC node and eliminate the weighted perturbations after secure aggregation. Notably, DRL-PBFL effectively achieves Byzantine robustness in FL even though the vehicles' data is not independently and identically distributed (non-IID). Experiments on three benchmark datasets with different adversarial settings show that DRL-PBFL effectively improves the global model accuracy under tailored Byzantine attacks compared with state-of-the-art schemes.
- 28 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Selected Areas in Communications
- Vol. 43 (6) , 1923-1937
Knowledge-Aware Privacy-Preserving Model Customization in Zero-Trust Federated Learning Model Marketplaces
Federated learning (FL) model marketplaces require qualified workers to collaboratively train customized models. However, recruiting optimal workers on a limited budget in non-independent and identically distributed (non-IID) data settings remains a fundamental issue. Moreover, inadequate quality verification exposes the marketplace to spoofing and poisoning attacks, while verifying data and model quality without accessing local storage remains a significant dilemma. To bridge the research gap, this paper proposes a knowledge-aware model customization scheme in FL model marketplaces, to facilitate zero-trust worker recruitment and verification while ensuring privacy preservation. Specifically, (i) we design a knowledge-aware quality evaluation mechanism by leveraging the knowledge of workers, i.e., soft-label predictions of their local models on a privacy-free reference dataset (provided by the customer), to assess their data quality in a privacy-preserving manner. (ii) We formulate the optimal worker recruitment problem under budget constraints as an NP-hard integer programming problem and design a dynamic programming-based optimal worker recruitment algorithm with budget feasibility and computational efficiency. (iii) We devise a two-stage zero-trust quality verification mechanism by utilizing zero-knowledge proof (ZKP) to exclude distrustful workers, thereby preventing spoofing and poisoning attacks. Extensive experimental results demonstrate that the proposed scheme enhances model customization performance by up to 34.3% on label-skewed non-IID data and 36.2% on feature-skewed non-IID data compared with existing representatives.
- 28 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Selected Areas in Communications
- Vol. 43 (6) , 1923-1937
Knowledge-Aware Privacy-Preserving Model Customization in Zero-Trust Federated Learning Model Marketplaces
Federated learning (FL) model marketplaces require qualified workers to collaboratively train customized models. However, recruiting optimal workers on a limited budget in non-independent and identically distributed (non-IID) data settings remains a fundamental issue. Moreover, inadequate quality verification exposes the marketplace to spoofing and poisoning attacks, while verifying data and model quality without accessing local storage remains a significant dilemma. To bridge the research gap, this paper proposes a knowledge-aware model customization scheme in FL model marketplaces, to facilitate zero-trust worker recruitment and verification while ensuring privacy preservation. Specifically, (i) we design a knowledge-aware quality evaluation mechanism by leveraging the knowledge of workers, i.e., soft-label predictions of their local models on a privacy-free reference dataset (provided by the customer), to assess their data quality in a privacy-preserving manner. (ii) We formulate the optimal worker recruitment problem under budget constraints as an NP-hard integer programming problem and design a dynamic programming-based optimal worker recruitment algorithm with budget feasibility and computational efficiency. (iii) We devise a two-stage zero-trust quality verification mechanism by utilizing zero-knowledge proof (ZKP) to exclude distrustful workers, thereby preventing spoofing and poisoning attacks. Extensive experimental results demonstrate that the proposed scheme enhances model customization performance by up to 34.3% on label-skewed non-IID data and 36.2% on feature-skewed non-IID data compared with existing representatives.