Existing generative image transformers follow a two-stage generation paradigm, where the first stage learns a codebook to encode images into discrete codes via vector quantization, and the second stage completes the image generation based on the learned codebook. However, existing methods ignore the naturally varying information densities across different image regions and indiscriminately encode fixed-size regions into fixed-length codes, resulting in insufficient encoding in important regions and redundant encoding in unimportant ones, which degrades both the image generation quality and speed. To address this challenge, we propose a novel information-density-based variable-length image coding and generation framework. In the first stage, our Dynamic Quantization VAE++ (DQVAE++) performs information-adaptive encoding by assigning variable-length codes to image regions according to their information densities, yielding more accurate and robust code representations. In the second stage, the Dynamic Generative Image Transformer (DGiT) enables information-adaptive image generation in both autoregressive and non-autoregressive manners. Specifically, for autoregressive (AR) generation, DGiT-AR generates images autoregressively from coarse-grained regions (smooth areas with fewer codes) to fine-grained regions (detailed areas with more codes). This is accomplished through a novel stacked-transformer architecture that alternately models the position and content of image codes, and a novel heterogeneous embedding scheme to distinguish codes of different granularities. Similarly, for non-autoregressive (NAR) generation, DGiT-NAR introduces a novel information-prioritized mask scheduling mechanism, prioritizing the generation of key structural regions with higher information density. This enables more coherent modeling of global structures initially, followed by a more effective synthesis of local details subsequently. Comprehensive experiments on unconditional and conditional image generation validate the superiority of our proposed variable-length coding in both effectiveness and efficiency.

Few-shot knowledge graph completion (few-shot KGC) mines unseen knowledge by leveraging meta-learning and contrastive learning to achieve accurate predictions with limited triples. Recent studies have focused on designing distance or similarity metrics to provide better knowledge representation between entities and relations. However, three issues with negative sampling remain unexplored: 1) the construction of negative queries heavily relies on manual experience in selecting candidate tail entities, 2) the constructed negative queries may mislabel potential true facts, and 3) the varying difficulties of negative queries are ignored. To solve the above issues, in this paper, we introduce curriculum learning into few-shot KGC and propose a novel few-shot KGC framework empowered by an adaptive negative sampling mechanism, which can eliminate the dependence on any additional manual experience, reduce mislabeling, and generate negative queries with appropriate difficulty. Specifically, the proposed framework includes two alternating phases. In the negative sampling phase, we first design a novel positive-unlabeled learning based scoring function with a type-related candidates encoder and then build a variable-speed sliding window based pacing function to select negative queries with appropriate learning difficulty under current training step. In the meta-training phase, we develop an adapted triple-oriented knowledge encoder to provide accurate representation for queries. Experimental results demonstrate that the proposed framework outperforms the state-of-the-art baselines and provides negative queries with appropriate difficulty in few-shot KGC.

Collaborative perception can significantly enhance the perceptual capabilities of autonomous vehicles by sharing sensing information through vehicular communications. However, large-scale sharing of sensing information often results in unsustainable network loads, making it challenging to maximize perception performance with limited communication resources in complex environments. To address this challenge, we propose FullPerception, an innovative cooperative perception framework that jointly orchestrates sensing information sharing and communication resource allocation at the network level. FullPerception advocates for the sharing of semantic information (neural network features) within critical areas, i.e., blind spots. With limited communication resources, FullPerception strategically eliminates these blind spots to maximize the accumulated perception performance. We formulate this strategy as a weighted optimization problem and prove its NP-hardness. We propose a simple yet effective algorithm, Proactive Conflict-free Scheduling (PCS), which guarantees a good performance ratio by considering broader contexts. PCS is meticulously combined with recursive structure, accounting for both the overall and future contexts to determine link scheduling and resource allocation. We demonstrate that FullPerception improves perception accuracy by 20% relative to single-vehicle systems and by 10% compared to existing scheduling methods through large-scale comprehensive joint simulation experiments.

Large language models are increasingly deployed across mobile and edge environments, where privacy-sensitive and heterogeneous user data raise critical concerns of copyright infringement, data leakage, and regulatory non-compliance. Ma chine unlearning has thus emerged as an essential capability to remove the influence of specific data without full retraining. However, two key challenges remain open: 1) how to quantify unlearning to enable data valuation without retraining, especially since the massive scale of pretraining makes it infeasible to evaluate the contribution of individual data samples in advance, and 2) how to verify the correctness without retraining to ensure that third-party auditors can efficiently confirm the complete removal of targeted data influence. To address the aforementioned challenges, in this paper, we design a dual-stage machine unlearning framework to quantify the contribution of forgotten data and certify data removal without full retraining, serving as an auditing layer for first-order unlearning methods. Specifically, we design a run-time Shapley value-based unlearned data evaluation mechanism that utilizes a first-order approximation strategy to estimate the marginal contribution of forgotten samples. Moreover, we propose a proof of unlearning mechanism that generates compact, auditable artifacts of the unlearning process to efficiently verify that the targeted data influence has been completely removed. Compared with five state-of-the-art unlearning baselines, our approach achieves effectiveness in data valuation, stronger guarantees of removal correctness, and lower computational overhead.

Aerial base station (AeBS) is a promising technology for providing wireless coverage to ground user equipment. Traditional methods of optimizing AeBS networks often rely on pre-known distribution models of ground user equipment. However, in practical scenarios such as natural disasters or temporary large-scale public events, the distribution of user clusters is often unknown, posing challenges for the deployment and application of AeBS. To adapt to complex and unknown user environments, this paper studies a method of estimating information from local to global and proposes a multi-agent AeBSs autonomous deployment algorithm based on deep reinforcement learning (DRL). This method attempts to dynamically deploy AeBS to autonomously identify hotspots by sensing user equipment signals without network-side assistance, providing a more comprehensive and intelligent solution for AeBS deployment. Simulation results indicate that our method effectively guides the autonomous deployment of AeBS in emergency scenarios, addressing the challenge of the lack of network-side assistance.

Multimodal event extraction task aims to identify event types and arguments from visual and textual representations related to events.Due to the high cost of multimedia training data, previous methods mainly focused on weakly alignment of excellent unimodal encoders.However, they ignore the conflict between event understanding and image recognition, resulting in redundant feature perception affecting the understanding of multimodal events.In this paper, we propose a multimodal event extraction strategy with a multi-level redundant feature selection mechanism, which enhances the event understanding ability of multimodal large language models by leveraging knowledge editing techniques, and requires no additional parameter optimization work.Extensive experiments show that our method outperforms the state-ofthe-art (SOTA) baselines on the M2E2 benchmark.Compared with the highest baseline, we achieve a 34% improvement of Precision on event extraction and a 11% improvement of F1 on argument extraction.

The large number of devices and complex communication requirements pose challenges to ensuring the security of Non-Terrestrial Internet of Things (NT-IoT). The large-scale data and complex communication requirements make accurate detection of malicious traffic even more challenging in NT-IoT. Hypergraph neural networks have strong performance in extracting multi-relational features. However, most existing hypergraph neural networks are tailored for graph data, and hyperedge construction methods are not well-suited. To address these challenges, we propose a malicious encrypted traffic detection method based on a hypergraph neural network. First, we propose an efficient hypergraph construction method for encrypted traffic named JointKNN. JointKNN calculates the Euclidean distance between traffic flows and adds the target nodes into the neighbor sets to form the hyperedges. Then, we propose an Encrypted Traffic HyperGraph Convolution Network (ETHGCN), which takes the encrypted traffic hypergraph as the input. ETHGCN extracts and fuses both connection and temporal features to accurately detect malicious traffic. We conduct comparative experiments on IoT and Onion Network encrypted traffic datasets for multi-class and binary classification tasks. Results indicate that ETHGCN achieves an accuracy exceeding 99.8% in IoT tasks and demonstrates an improvement of nearly 20% in Onion Network tasks.

Unifying diverse image generation tasks within a single framework remains a fundamental challenge in visual generation. While large language models (LLMs) achieve unification through task-agnostic data and generation, existing visual generation models fail to meet these principles. Current approaches either rely on per-task datasets and large-scale training or adapt pre-trained image models with task-specific modifications, limiting their generalizability. In this work, we explore video models as a foundation for unified image generation, leveraging their inherent ability to model temporal correlations. We introduce RealGeneral, a novel framework that reformulates image generation as a conditional frame prediction task, analogous to in-context learning in LLMs. To bridge the gap between video models and condition-image pairs, we propose (1) a Unified Conditional Embedding module for multi-modal alignment and (2) a Unified Stream DiT Block with decoupled adaptive LayerNorm and attention mask to mitigate cross-modal interference. RealGeneral demonstrates effectiveness in multiple important visual generation tasks, e.g., it achieves a 14.5% improvement in subject similarity for customized generation and a 10% enhancement in image quality for canny-to-image task. Project page: https://lyne1.github.io/RealGeneral/

Federated Learning has garnered significant attention in practical applications due to its privacy-preserving properties but faces serious threats from backdoor attacks. Current defenses primarily rely on server-side anomaly detection and robust aggregation, but lack systematic strategies for proactively erasing backdoors from the perspective of attackers. To bridge this gap, we propose an efficient and stealthy federated backdoor removal framework. Specifically, our method incorporates a two-stage training approach: reinforced negative learning and positive memory recovery. In addition, we intro-duce a hybrid regularization strategy that combines dynamic L1 regularization with Elastic Weight Consolidation, together with a synchronized differential amplification mechanism for both weights and buffers and a global norm clipping strategy. These components collectively effectively erase backdoor effects, maintain main-task accuracy, and significantly reduce detection risk on the server side. Experimental evaluations demonstrate that, compared to existing backdoor unlearning methods, our approach decreases the success rate of the backdoor attack by up to 11% without compromising the primary precision. Furthermore, our method substantially improves the stealthiness of the update, reducing the L2 norm fluctuations to less than 33% of the baseline levels.

Blockchain-based covert communication (BCC) provides high reliability and anonymity by embedding secret data into blockchain transactions. However, existing BCC approaches still face three fundamental limitations: (i) data mining risk, since transactions containing the secret data are permanently recorded on-chain and may be detected perpetually; (ii) limited efficiency, as only small payloads (e.g., 256 bits) can be carried per transaction; and (iii) private key leakage, where receivers often need access to the sender’s private key and may incur private key exposure. To address these issues, we propose a novel covert communication model with transaction withdrawal (BCC-TW) and a multisig-based data embedding scheme (MUL-DE). BCC-TW prevents covert transactions from being confirmed by constructing higher-fee double-spend transactions, thereby ensuring that secret data only exists temporarily in the mempool. MUL-DE encodes data into redundant public keys of Bitcoin multisig addresses, thus enabling higher efficiency and avoiding private key exposure. We implement a prototype on Bitcoin testnet and evaluate its concealment and efficiency. Experimental results demonstrate that the proposed approach achieves strong indistinguishability against statistical and deep-learning-based detectors, improves communication efficiency up to 251 bits per public key, and significantly reduces cost compared with state-of-the-art baselines.

As the development of B5G and 6G continues to progress, higher network bandwidth and increasingly complex vehicle connectivity are driving greater concurrency in highly dynamic and delay-sensitive transportation tasks within the Internet of Vehicles (IoV). Existing resource allocation methods such as Deep Reinforcement Learning (DRL), Graph Neural Network (GNN), Lyapunov and simple Transformer series often result in insufficient individual consideration or unprioritized attention on key resource characteristics, causing high task execution time cost and energy consumption. To overcome above problems, this paper proposes a Dynamic Self-Feedback (DSF) resource allocation approach. First, DSF models task latency and requirements along with diverse computing power to support allocation and dynamically adjusts the dimensions of self-attention heads according to resource consumption prediction in a self-feedback manner. Then, DSF adjusts dimensions of attention embedding to light and heavy tasks as feedback and leads next round of allocation optimization. Therefore, DSF enables individually tailored and energy-efficient allocation of computing resources for high concurrent IoV tasks. Simulations show the proposed mechanism achieves up to 85% tasks execution efficiency and 38% fewer timeout tasks, more than 50% of low energy consumption tasks after allocation, with almost 100% units having a workload lower than 40%.

With the rapid development of Internet of Things (IoT) technology, an increasing number of resource-constrained devices operate in dynamic and heterogeneous network environments, posing challenges for efficient image transmission. Multi-user semantic communication (SC) enables reduced bandwidth consumption and enhanced noise resilience by understanding the intrinsic meaning of information and sharing common semantic features across devices, offering great potential for widespread applications in various IoT scenarios. However, current multi-users SC approaches for image transmission lack adaptability and fail to consider both content and style features, leading to degraded image reconstruction quality. Moreover, semantic redundancy among devices remains underutilized, limiting bandwidth efficiency in IoT networks. To address these limitations, in this paper, a novel multi-user content-style adaptive semantic communication system for image transmission in IoT scenarios is proposed. Specifically, a dual-branch semantic information extraction and adaptive recovery scheme is first established, which simultaneously captures and adaptively fuses semantic content and style features to improve reconstruction quality. Secondly, an adaptive common information extraction and enhanced coding module is introduced for resource-limited IoT devices, which dynamically adjusts the transmission rate based on varying channel conditions and the computational capabilities of different users, further optimizing communication performance. Finally, experimental results show that the proposed method improves peak signal-to-noise (PSNR) by at least 10% under poor SNR conditions for multi-users semantic communication, compared to baseline methods.

In this paper, we present$\mathtt {Antelope}$, a semi-honest large-scale secure inference system without revealing either clients’ data or model parameters. The main contributions of$\mathtt {Antelope}$are new two-party computation (2PC) protocols over a ring$\mathbb {Z}_{2^\ell }$for non-linear layers, which optimize the online computation and communication overhead thus outperforming the state-of-the-art 2PC systems. Specifically, we reformulate the comparison function as an Equality-to-Zero test followed by multiplication, decoupling the bit-wise rounding dependency in traditional secret sharing-based bit extraction. With this technique, the evaluation of the ReLU non-linear activation function is$1.7\times$-$84.5\times$faster than existing solutions in online communication cost. We also develop a suite of optimizations that improve the efficiency of secure division protocols, which are tailored to different divisor settings in the neural networks. We extend our protocols to construct efficient implementations for several building blocks such as ReLU, Maxpool, truncation, and Softmax. End-to-end evaluation on realistic ImageNet-scale networks demonstrates that$\mathtt {Antelope}$achieves over$22.3\times$and$23.0\times$online runtime speedups in LAN and WAN settings, respectively, without accuracy loss, compared to the state-of-the-art works.

Maritime surveillance plays a vital role in reducing maritime accidents and improving maritime safety. To enhance situational awareness for maritime movements, deep learning-based visual object detection has become an important part of maritime surveillance. However, the detection results are highly dependent on the training datasets collected from different departments (i.e., clients). If the sub-datasets from departments are sensitive and private in cross-department maritime surveillance, it will be intractable to directly combine these sub-datasets to train the learning-based object detection method. To solve this issue, we propose a federated learning-driven cross-spatial vessel detection model, called FLCSDet, for maritime surveillance with privacy preservation. In particular, an efficient multi-scale attention module is integrated into our FLCSDet to achieve local cross-spatial feature learning. To improve the federated-learning aggregation method, we propose an optimized algorithm based on the proportion of valid data on departments to adaptively select the allocating weights and preserve the specific characteristics of client data. In addition, we employ transfer learning to further improve the robustness and convergence of our FLCSDet under different experimental scenarios. Compared with several representative federated learning-based detection methods, our FLCSDet could achieve superior detection performance in terms of both quantitative and qualitative results. Moreover, comprehensive experiments conducted on real datasets from both inland waterways and open seas demonstrate the robustness and generalization of our method in intelligent transportation systems. The source code is available athttps://github.com/huangyanh/FLCSDet.

Federated Learning (FL) is a privacy-preserving computing paradigm that enables participants to collaboratively train a global model without exchanging their raw personal data. Due to frequent communication and data heterogeneity of devices with unique local data distributions, FL faces a significant issue with slow convergence speed. To achieve fast convergence, existing methods adjust hyperparameters in FL to reduce the volume of model updates, the number of participating devices, and local iterations. However, most focus on only part of the hyperparameters and primarily rely on analytical optimization. A more integrated and dynamic coordination of all hyperparameters is needed. To address this issue, we first propose an efficient FL framework enabled by rand-m sparsification and stochastic quantization methods. For this framework, we conduct a rigorous theoretical analysis to explore the trade-offs among quantization level, sparsification level, device participation, and local iteration. To improve convergence speed, we also design a Deep Reinforcement Learning (DRL)-based strategy to dynamically coordinate these hyperparameters. Experimental results show that our method can improve convergence speed by at least 8% compared to the existing approaches.

Dataset distillation is utilized to condense large datasets into smaller synthetic counterparts, effectively reducing their size while preserving their crucial characteristics. In Federated Learning (FL) scenarios, where individual devices or servers often lack substantial computational power or storage capacity, the use of dataset distillation becomes particularly advantageous for processing large volumes of data efficiently. Current research in dataset distillation for FL has primarily focused on enhancing accuracy and reducing communication complexity, but it has largely neglected the potential risk of backdoor attacks. To solve this issue, in this paper, we propose three adaptive dataset condensation based backdoor attacks against dataset distillation for FL. Adaptive attacks in dataset distillation for FL dynamically modify triggers during the training process. These triggers, embedded in the synthetic data, are designed to bypass traditional security detection. Moreover, these attacks employ self-adaptive perturbations to effectively respond to variations in the model's parameters. Experimental results show that the proposed adaptive attacks achieve at least 5.87% higher success rates, while maintaining almost the same clean test accuracy, compared to three benchmark methods.

AI-generated content (AIGC) gained immense popularity across various domains, retrieving valuable training data by using free APIs (Application Programming Interfaces) from various applications and utilizing AI techniques to generate content automatically. However, concerns have been raised regarding unfair payment for the utilization of valuable training data between data owners and AIGC services providers (ASPs). Blockchain oracle can establish trust between them and bridge on-chain and off-chain training data trading. However, the integration of blockchain and AIGC services is challenged by the scalability of oracle consensus. It is essential not only to support a high volume of data requests from ASPs but also to ensure timely and accurate training data responses. To solve the above issues, we first propose an API-based decentralized AIGC data sharing framework and introduce a blockchain oracle to help ASPs retrieve training data from off-chain data owners. We then design flooding-based oracle consensus protocols to achieve scalable and efficient interactions between AIGC and data owners. Theoretical analysis and simulation results demonstrate that the proposed mechanism can significantly reduce communication overheads.

With the advancement of federated crowdsourcing services, the associated privacy concerns have attracted growing attention from both academia and industry. Existing privacy laws impose strict requirements concerning the right to be forgotten for data used in training AI models. In federated crowdsourcing services, the right to be forgotten is guaranteed through federated unlearning. Current federated unlearning solutions encompass a two-step process: first, eliminating model updates associated with the target data to achieve unlearning, followed by retraining among the remaining clients to restore the performance of federated crowdsourcing services. However, this indiscriminate removal of model updates, while safeguarding the privacy of the target data, also greatly undermines the generalization performance of the global model. Moreover, relying on client-side retraining imposes additional economic costs on the federated crowdsourcing service. To tackle the above issues, this paper proposes an efficient federated unlearning framework for federated crowdsourcing services, which is based on adaptive parameter clipping and data-free distillation. We first compute the Fisher information matrix (FIM) to approximate the correlation between the target data and all model parameters, which is utilized to adaptively clip each parameter of the global model. Then, we model the softmax layer of the global model to synthesize pseudo-samples, enabling the retrain process on the crowdsourcing platform for the recovery of generalization performance. We conducted extensive experiments on three datasets, and the results demonstrate that our proposed framework not only possesses outstanding data removal capability but also outperforms the comparison methods in terms of computation time and storage space.

Optimizing various wireless user tasks poses a significant challenge for networking systems because of the expanding range of user requirements. Despite advancements in Deep Reinforcement Learning (DRL), the need for customized optimization tasks for individual users complicates developing and applying numerous DRL models, leading to substantial computation resource and energy consumption and can lead to inconsistent outcomes. To address this issue, we propose a novel approach utilizing a Mixture of Experts (MoE) framework, augmented with Large Language Models (LLMs), to analyze user objectives and constraints effectively, select specialized DRL experts, and weigh each decision from the participating experts. Specifically, we develop a gate network to oversee the expert models, allowing a collective of experts to tackle a wide array of new tasks. Furthermore, we innovatively substitute the traditional gate network with an LLM, leveraging its advanced reasoning capabilities to manage expert model selection for joint decisions. Our proposed method reduces the need to train new DRL models for each unique optimization problem, decreasing energy consumption and AI model implementation costs. The LLMenabled MoE approach is validated through a general maze navigation task and a specific network service provider utility maximization task, demonstrating its effectiveness and practical applicability in optimizing complex networking systems.

With the development of AI-Generated Content (AIGC), data is becoming increasingly important, while the right of data to be forgotten, which is defined in the General Data Protection Regulation (GDPR) and permits data owners to remove information from AIGC models, is also arising. To protect this right in a distributed manner corresponding to federated learning, federated unlearning is employed to eliminate history model updates and unlearn the global model to mitigate data effects from the targeted clients intending to withdraw from training tasks. To diminish centralization failures, the hierarchical federated framework that is distributed and collaborative can be integrated into the unlearning process, wherein each cluster can support multiple AIGC tasks. However, two issues remain unexplored in current federated unlearning solutions: 1) getting remaining clients, those not withdraw from the task, to join the unlearning process, which demands additional resources and notably has fewer benefits than federated learning, particularly in achieving the original performance via alternative unlearning processes and 2) exploring mechanisms for dynamic unlearning in the selection of remaining clients possessing unbalanced data to avoid starting the unlearning from scratch. We initially consider a two-level incentive and unlearning mechanism to address the aforementioned challenges. At the lower level, we utilize evolutionary game theory to model the dynamic participation process, aiming to attract remaining clients to participate in retraining tasks. At the upper level, we integrate deep reinforcement learning into federated unlearning to dynamically select remaining clients to join the unlearning process to mitigate the bias introduced by the unbalanced data distribution among clients. Experimental results demonstrate that the proposed mechanisms outperform comparative methods, enhancing utilities and improving accuracy.

AI-Generated Content (AIGC) services are essential in developing the Metaverse, providing various digital content to build shared virtual environments. The services can also offer personalized content with user assistance, making the Metaverse more human-centric. However, user-assisted content creation requires significant communication resources to exchange data and construct trust among unknown Metaverse participants, which challenges the traditional centralized communication paradigm. To address the above challenge, we integrate blockchain with semantic communication to establish decentralized trust among participants, reducing communication overhead and improving trustworthiness for AIGC services in Metaverse. To solve the out-of-distribution issue in data provided by users, we utilize the invariant risk minimization method to extract invariant semantic information across multiple virtual environments. To guarantee trustworthiness of digital contents, we also design a smart contract-based verification mechanism to prevent random outcomes of AIGC services. We utilize semantic information and quality of digital contents provided by the above mechanisms as metrics to develop a Stackelberg game-based content caching mechanism, which can maximize the profits of Metaverse participants. Simulation results show that the proposed semantic extraction and caching mechanism can improve accuracy by almost 15% and utility by 30% compared to other mechanisms.

Semantic communication constructs a promising and lightweight paradigm for participants to transmit semantic information to each other. However, it suffers from untrust among participants, insecure transmission, and underestimation of the value of semantic information. Blockchain is a decentralized peer-to-peer network that can provide participants with transparent, secure, and trusting environments to implement data sharing. The integration of blockchain and semantic communication is considered a promising paradigm for overcoming the above challenges. However, a unified integration framework has not been studied. Thus, in this article, we first propose a novel blockchain and semantic ecosystems-based framework to share semantic information. We also design the proof of semantic mechanism to solve the garbage-in garbage-out challenge of blockchain. Moreover, we construct a state channel and task-relevant information bottleneck approach-based semantic sharing mechanism to improve the efficiency of semantic sharing. Simulation results show that the proposed mechanisms are verifiable and efficient, and perform better than the compared methods.

AI-Generated Content (AIGC) services are essential in developing the Metaverse, providing various digital content to build shared virtual environments. AIGC services can offer personalized content with user assistance, making the Metaverse more human-centric. However, it is difficult for participants to exchange data and construct trust among unknown Metaverse participants. To address the above challenge, we propose an integration of blockchain and AIGC to construct decentralized trust among participants. We design a smart contract-based verification mechanism to prevent random outcomes of AIGC services and guarantee the authenticity of digital contents. Given the quality of digital contents provided by the previous mechanisms, we then utilize them as metrics to establish a Stackelberg game-based content caching mechanism to maximize Metaverse participants’ profits. Simulation results show that the proposed caching mechanism can improve utility by 30% compared to other mechanisms.

Federated Learning (FL) is a distributed learning framework that enables data sharing among multiple devices to protect data privacy. Blockchain is a decentralized ledger that can record data securely and reliably. The blockchain-based FL (BFL) framework has been used to share data and computing resources in multiple clusters. However, for the BFL framework among multi-institutional clusters, data sparsity in a cluster is a key issue. Most of the relevant works assume that the data in one cluster is rich enough to build a suitable model, which is not always satisfied in all scenarios. One method to address the problem is that enlarging the size of a BFL cluster that covers as many nodes as possible is one way. However, this method will increase communication overheads and reduce transaction throughput of the blockchain. To solve the above issues, we propose a communication-efficient blockchain-based FL framework called CCFL, which connects multiple BFL clusters to solve the data sparsity issue. We also design a pearson correlation coefficient-based dynamic model filtering mechanism that filters unnecessary models to reduce communication costs and exclude malicious models. Moreover, we illustrate a reliable contribution-based interactive validation reputation mechanism to prevent malicious nodes from participating in the training. We carry out some experiments to show the feasibility and efficiency of the proposed framework.

Blockchain-based Federated learning, called BFL, has attracted widespread attention to construct trust among multiple parties and solve a single point of failure of the central server while protecting privacy. Many researches utilize cluster and cross-chain technologies to improve poor model quality and interoperability between clusters. However, those researches still suffer from 1) high communication overhead when devices of clusters locate far away, and 2) high consensus latency since devices require frequent interactions on consensus. In this paper, we propose a cross-cluster federated learning framework based on state channels, called SCFL, to split devices into multiple clusters according to locations. We also propose a cross-cluster consensus algorithm based on cross-chain and state channels to improve the security and efficiency of off-chain and inter-chain interactions. And we also propose a hierarchical clustering method to make the model adaptable to the partition scenarios where the data is non-IID. Numerical results show that SCFL can effectively solve data sparse problems and improve the system efficiency in non-IID data partitioning cases.

The rapid development of digital technology is reshaping the architecture of the Industrial Internet of Things (IIoT). The traditional architecture cannot process vast amounts of data exchanges and provide entities with trust. The future IIoT is expected to be a decentralized architecture in which blockchain and digital twin-driven IIoT can enable trusted data exchanges. However, this architecture cannot obtain huge amounts of external real-time data and isolated data. Moreover, it cannot handle complex industrial computing tasks. Therefore, we combine oracle with decentralized learning to propose a novel IIoT-oriented digital twin architecture. We also propose an effective decentralized collaboration mechanism to support external data and resources exchanges. Moreover, we propose a novel computing collaboration mechanism to expand the learning capabilities of the industrial ecology. Experiments show that our proposed paradigm has less processing time, a more stable process, and better learning ability compared to other paradigms.

Blockchain-based Federated Learning (FL) technology enables vehicles to make smart decisions, improving vehicular services and enhancing the driving experience through a secure and privacy-preserving manner in Intelligent Transportation Systems (ITS). Many existing works exploit two-layer blockchain-based FL frameworks consisting of a mainchain and subchains for data interactions among intelligent vehicles, which resolve the limited throughput issue of single blockchain-based vehicular networks. However, the existing two-layer frameworks still suffer from a) strong dependency on predetermined and fixed parameters of vehicular blockchains which limit blockchain throughput and reliability; and b) high communication costs incurred by interactions among intelligent vehicles between the mainchain and subchains. To address the above challenges, we first design an adaptive blockchain-enabled FL framework for ITS based on blockchain sharding to facilitate decentralized vehicular data flows among intelligent vehicles. A streamline-based shard transmission mechanism is proposed to ensure communication efficiency almost without compromising the FL accuracy. We further formulate the proposed framework and propose an adaptive sharding mechanism using Deep Reinforcement Learning to automate the selection of parameters of vehicular shards. Numerical results clearly show that the proposed framework and mechanisms achieve adaptive, communication-efficient, credible, and scalable data interactions among intelligent vehicles.

Blockchain as a new decentralized chain structure can be used in edge computing to solve the security issue caused by edge nodes in model training. However, large amounts of data exchanges in the process of model training of edge computing reduce the performance of blockchain, and meanwhile, the block needed to be saved in the edge node challenges storage capacity of the edge node. Therefore, in the paper we propose a safe and efficient model training mechanism based on onchain and offchain collaboration. In the mechanism, edge nodes train models locally, store the model parameters in offchain and only return identifiers for model aggregation. By the method, the storage pressure of the edge node is reduced and the efficiency of executing consensus algorithms are increased. Moreover, in the mechanism we design a reputation evaluation model based on confidence factors to avoid the uploading of random and wrong data of edge nodes. Evaluation results show that our schemes can reduce the average delay and resources consumption, increase transaction throughput and maintain security compared with a state-of-the-art scheme.

Federated learning (FL) is a distributed framework for machine learning (ML) model training. Training agents upload local model parameters rather than original training data, and the central server performs parameter aggregation. FL can protect user data privacy and break the information island when training the ML model. Federated Average (FedAvg) is an aggregation method commonly used in the FL training task. The central server calculates the mean value of the local model parameters to obtain the new global parameters. FedAvg assumes that all the training agents are honest, which means the central server lacks terminal agents' knowability. When there are attackers in the training agents, the global model's performance may be deeply affected, and the training task cannot be completed normally. To solve this problem, we propose a Federated Learning method with aggregation based on the Importance of training agent (FedIM), in which the central server performs pre-evaluation on the agent parameters before aggregation, calculates the weights of parameters according to the historical behavior records of training terminals and performs federated aggregation to improve the anti-poisoning ability of learning task. Experiments show that our method can effectively improve the global model's anti-poisoning ability and accelerate the training speed compared with the FedAvg method when malicious agents are involved.

The blockchain system allows various trans-actions and information storage to be executed in a decentralized manner, while smart contracts require multiple nodes to be executed in the local sandbox environment according to preset settings to ensure the consistency of each node, which makes smart contracts unable to proactively obtain data from the outside world. Decentralized oracle can realize the acquisition of off-chain data with a low speed under the premise of ensuring the decentralization of the blockchain. Some oracles use on-chain data storage and maintenance to speed up data acquisition, but this will face higher costs of data storage and maintenance, so current oracles cannot simultaneously ensure privacy and security while taking into account execution cost and processing speed. In this article, we propose Select-Storage, a new oracle design pattern to achieve low operating cost and high processing speed without compromising security. Through experimental analysis, and comparison with other design patterns in processing time and on-chain and off-chain call costs, we have proved the superiority of the Select-Storage design pattern.