Large models, renowned for superior performance, outperform smaller ones even without billion-parameter scales. While mobile network servers have ample computational resources to support larger models than client devices, privacy constraints prevent clients from directly sharing their raw data. Federated Learning (FL) enables decentralized clients to collaboratively train a shared model by exchanging model parameters instead of transmitting raw data. Yet, it requires a uniform model architecture and multiple communication rounds, which neglect resource heterogeneity, impose heavy computational demands on clients, and increase communication overhead. To address these challenges, we propose FedOL, to construct a larger and more comprehensive server model in one-shot settings (i.e., in a single communication round). Instead of model parameter sharing, FedOL employs knowledge distillation, where clients only exchange model prediction outputs on an unlabeled public dataset. This reduces communication overhead by transmitting compact predictions instead of full model weights and enables model customization by allowing heterogeneous model architectures. A key challenge in this setting is that client predictions may be biased due to skewed local data distributions, and the lack of ground-truth labels in the public dataset further complicates reliable learning. To mitigate these issues, FedOL introduces a specialized objective function that iteratively refines pseudo-labels and the server model, improving learning reliability. To complement this, FedOL incorporates a tailored pseudo-label generation and knowledge distillation strategy that effectively integrates diverse knowledge. Simulation results show that FedOL significantly outperforms existing baselines, offering a cost-effective solution for mobile networks where clients possess valuable private data but limited computational resources.

Native AI support is a key objective in the evolution of 6G networks, with Federated Learning (FL) emerging as a promising paradigm. FL allows decentralized clients to collaboratively train an AI model without directly sharing their data, preserving privacy. Clients train local models on private data and share model updates, which a central server aggregates to refine the global model and redistribute it for the next iteration. However, client data heterogeneity slows convergence and reduces model accuracy, and frequent client participation imposes communication and computational burdens. To address these challenges, we propose FedABC, an innovative client selection algorithm designed to take a long-term view in managing data heterogeneity and optimizing client participation. Inspired by attention mechanisms, FedABC prioritizes informative clients by evaluating both model similarity and each model's unique contributions to the global model. Moreover, considering the evolving demands of the global model, we formulate an optimization problem to guide FedABC throughout the training process. Following the “later-is-better” principle, FedABC adaptively adjusts the client selection threshold, encouraging greater participation in later training stages. Extensive simulations on CIFAR-10 demonstrate that FedABC significantly outperforms existing approaches in model accuracy and client participation efficiency, achieving comparable performance with 32% fewer clients than the classical FL algorithm FedAvg, and 3.5% higher accuracy with 2% fewer clients than the state-of-the-art. This work marks a step toward deploying FL in heterogeneous, resource-constrained environments, thereby supporting native AI capabilities in 6G networks.

Existing approaches defend against backdoor attacks in federated learning (FL) mainly through a) mitigating the impact of infected models, or b) excluding infected models. The former negatively impacts model accuracy, while the latter usually relies on globally clear boundaries between benign and infected model updates. However, in reality, model updates can easily become mixed and scattered throughout due to the diverse distributions of local data. This work focuses on excluding infected models in FL. Unlike previous perspectives from a global view, we propose Snowball, a novel anti-backdoor FL framework through bidirectional elections from an individual perspective inspired by one principle deduced by us and two principles in FL and deep learning. It is characterized by a) bottom-up election, where each candidate model update votes to several peer ones such that a few model updates are elected as selectees for aggregation; and b) top-down election, where selectees progressively enlarge themselves through picking up from the candidates. We compare Snowball with state-of-the-art defenses to backdoor attacks in FL on five real-world datasets, demonstrating its superior resistance to backdoor attacks and slight impact on the accuracy of the global model.

With the emergence of the sixth-generation (6G) communication technologies, the Internet of Vehicles (IoV) is rapidly developing with the coordination between intelligent networked vehicles, road infrastructures, and the cloud. However, the openness and dynamic nature of the IoV raise significant security and privacy concerns, highlighting the need for efficient authentication schemes. Conventional authentication schemes are no longer suitable for 6G-enabled IoV due to high latency, single point of failure, and heavy management costs. Additionally, existing literature on multi-domain authentication mainly investigates vehicle mobility, ignoring the challenges posed by vehicle heterogeneity. To fill this gap, we propose a multi-domain authentication scheme with conditional privacy preservation (MACPP) that considers administrative domains (AD) and geographic domains (GD) in the IoV. In MACPP, we design a novel identity-based signature scheme without requiring bilinear pairing for efficient authentication. Additionally, we propose a blockchain-assisted pseudonym management scheme (BAPM) to further improve system security by designing a dynamical sparse Merkle tree structure (DSMT). We demonstrate that the proposed MACPP satisfies the security requirements through an in-depth security analysis. Moreover, the experimental results demonstrate the effectiveness and efficiency of both MACPP and BAPM.

The property of Internet of Things (IoT) applications is their capability to execute tasks through the collaboration of interconnected IoT objects. However, IoT collaborations face significant challenges due to security threats that undermine their reliability. An uncertified task publisher may deceive IoT devices into executing illegal tasks, while malicious attackers may intercept and modify transmitted data. Existing works on IoT trusted management issues tend to concentrate on individual aspects, such as authentication, privacy protection, and access control. However, trusted management for IoT collaboration is a multifaceted and intricate endeavor that necessitates a comprehensive approach. To fill this gap, we propose a lightweight authentication-driven trusted management framework that includes a novel authentication and key agreement scheme to guarantee the validity of task publishers, with greatly reduced overheads compared to recent works. The framework also incorporates a distributed data storage scheme and a fine-grained access control mechanism. We record the interactive messages on the blockchain to ensure behavior traceability. We evaluate the authentication scheme through comparative experiments and formal security analysis, demonstrating its efficiency and effectiveness. The experimental results of data storage and acquisition in real-world IoT environments indicate that the proposed framework is a feasible solution for reliable IoT collaboration.

Multi-server jobs that request multiple computing resources and hold onto them during their execution dominate modern computing clusters. When allocating the multi-type resources to several co-located multi-server jobs simultaneously in online settings, it is difficult to make the tradeoff between the parallel computation gain and the internal communication overhead, apart from the resource contention between jobs. To study the computation-communication tradeoff, we model the computation gain as the speedup on the job completion time when it is executed in parallelism on multiple computing instances, and fit it with utilities of different concavities. Meanwhile, we take the dominant communication overhead as the penalty to be subtracted. To achieve a better gain-overhead tradeoff, we formulate an cumulative reward maximization program and design an online algorithm, namedOgaSched, to schedule multi-server jobs.OgaSchedallocates the multi-type resources to each arrived job in the ascending direction of the reward gradients. It has several parallel sub-procedures to accelerate its computation, which greatly reduces the complexity. We proved that it has a sublinear regret with general concave rewards. We also conduct extensive trace-driven simulations to validate the performance ofOgaSched. The results demonstrate thatOgaSchedoutperforms widely used heuristics by 11.33%, 7.75%, 13.89%, and 13.44%, respectively.

Integrating native AI support into the network architecture is an essential objective of 6G. Federated Learning (FL) emerges as a potential paradigm, facilitating decentralized AI model training across a diverse range of devices under the co-ordination of a central server. However, several challenges hinder its wide application in the 6G context, such as malicious attacks and privacy snooping on local model updates, and centralization pitfalls. This work proposes a trusted architecture for supporting FL, which utilizes Distributed Ledger Technology (DLT) and Graph Neural Network (GNN), including three key features. First, a pre-processing layer employing homomorphic encryption is incorporated to securely aggregate local models, preserving the privacy of individual models. Second, given the distributed nature and graph structure between clients and nodes in the pre-processing layer, GNN is leveraged to identify abnormal local models, enhancing system security. Third, DLT is utilized to decentralize the system by selecting one of the candidates to perform the central server's functions. Additionally, DLT ensures reliable data management by recording data exchanges in an immutable and transparent ledger. The feasibility of the novel architecture is validated through simulations, demonstrating improved performance in anomalous model detection and global model accuracy compared to relevant baselines.

In cross-silo federated learning (FL), the data among clients are usually statistically heterogeneous (aka not independent and identically distributed, non-IID) due to diversified data sources, lowering the accuracy of FL. Although many personalized FL (PFL) approaches have been proposed to address this issue, they are only suitable for data with specific degrees of statistical heterogeneity. In the real world, the heterogeneity of data among clients is often immeasurable due to privacy concern, making the targeted selection of PFL approaches difficult. Besides, in cross-silo FL, clients are usually from different organizations, tending to hold architecturally different private models. In this work, we propose a novel FL framework, FedAPEN, which combines mutual learning and ensemble learning to take the advantages of private and shared global models while allowing heterogeneous models. Within FedAPEN, we propose two mechanisms to coordinate and promote model ensemble such that FedAPEN achieves excellent accuracy on various data distributions without prior knowledge of data heterogeneity, and thus, obtains the adaptability to data heterogeneity. We conduct extensive experiments on four real-world datasets, including: 1) Fashion MNIST, CIFAR-10, and CIFAR-100, each with ten different types and degrees of label distribution skew; and 2) eICU with feature distribution skew. The experiments demonstrate that FedAPEN almost obtains superior accuracy on data with varying types and degrees of heterogeneity compared with baselines.

With the continuous functional enhancement of network services, a service usually presents a directed acyclic graphic (DAG) structure. This paper models the DAG task scheduling problem as a multi-objective optimization problem to balance the task execution efficiency, network traffic, and system load balance in componentized task deployment. To produce an instant decision, we propose the Cloud-edge Collaborative Task Scheduling (CCTS) Algorithm based on hybrid reward architecture deep reinforcement learning (DRL). Specifically, to reduce the redundancy of the state space of the Markov decision process, we use directed graph convolution networks and graph convolution networks (GCN) to embed the directed task graph and undirected network graph, respectively. Simulation results show that the proposed method outperforms the compared convolutional neural networks and GCN-based DRL schemes in reducing the system latency, energy cost, network traffic, and load balance.

AI/ML is envisioned to play an essential role in 6G mobile communication systems. The privacy-preserving capabil-ities of Federated Learning (FL) make it promising in vertical applications; however, the central server-based system and lack of trusted data management limit its widespread use. To effectively support FL as a service from a network architecture perspective, this work provides a comprehensive design including three key features: First, the network architecture enables transparent and traceable data management based on Distributed Ledger Technology (DLT) platform, and realizes distributed and off-chain data storage by adopting Distributed Data Storage Entity (DDSE). Second, the central aggregator of an FL service is decoupled from the data management scheme mentioned above, and is decentralized through smart contracts for aggregator selection among a set of aggregator candidates, with the selected aggregator subsequently responsible for client selection and model aggregation. Third, a completed set of procedures for FL services operations is defined. A simulation system is developed to verify the feasibility of the proposed architecture and to study the impact of introducing the data management mechanisms on the overall performance overhead. The results show that the impact is related to the FL settings, with a worst-case time overhead of 15% observed in selected test cases, i.e., 15% of the total time spent on the interactions with the DLT platform and DDSE.

Fine-grained service offloading in collaborative edge computing can make full use of the limited resource of edge nodes to achieve efficient parallel computing. It is imperative to select appropriate edge nodes for the subtask offloading in order to ensure the network’s load balance. However, there is a lack of research on computing offloading of end-to-end fine-grained services, and existing node selection algorithms can only be used in small-scale scenarios or networks with a fixed number of nodes. In this paper, we construct an end-to-end fine-grained computing offloading model, with load balancing as the optimization goal. Especially, a deep graph matching method, based on graph neural networks, is used for offloading node selection. It can be applied to dynamic and large-scale scenarios with strong generalization capability and fast execution speed. Compared with baseline algorithms, it greatly reduces the network load imbalance degree while ensuring a high acceptance ratio of services and meeting delay, location and resource constraints.

Knowledge graph completion (KGC) aims to perform link prediction to fill lost relations between entities by knowledge graph embedding (KGE). Translating embedding, as an efficient embedding method in KGE, is widely applied in numerous recent KGC models. However, these translating models may lack the ability to express various relation patterns and mapping properties for knowledge graphs (KGs). In this paper, a simple and well-performed translating model named TransC is proposed to express different relations. A multi-channel mechanism is defined firstly to constrain translating embeddings. Then a relation-aware transfer function is designed to break the expressive restriction and map triplets involving the same relation into a corresponding plane. We also mathematically prove that TransC is capable of expressing four popular relation patterns and all mapping properties. Finally, experimental results illustrate that TransC can efficiently represent the different relation patterns and properties and achieve better performance than state-of-the-art translating models.