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.

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.

Lightweight models with high accuracy is critical for edge intelligence. Although the Knowledge Distillation (KD) has been successfully applied to reduce the accuracy loss of quantized neural networks, especially for resource-constrained edge devices, the process of pre-training complex high-precision teacher networks in KD however, will bring huge training overhead. Recently proposed online distillation frameworks offer a good solution for teacher-free distillation, but the regularization effect and simple average aggregation of KD further weaken the representation capability of quantized models that have been reconstructed. In this work, we propose Precision-Mixed and Weight-Average Ensemble (PMWAE) consisting of multiple group members and a group leader. PMWAE provides additional knowledge by changing the bit-precision of the activation and generates aggregated weights for each member in group by attention-based mechanism. The ensemble knowledge is further passed to the group leader to obtain the final model. Extensive experiments on the CIFAR-10/100 and ImageNet-1K datasets show that our method outperforms the existing state-of-the-art methods, both on standard convolutions and depth-wise separable convolutions.

Federated learning (FL) lately has shown much promise in improving the shared model and preserving data privacy. However, these existing methods are only of limited utility in the Internet of Things (IoT) scenarios, as they either heavily depend on high-quality labeled data or only perform well under idealized conditions, which typically cannot be found in practical applications. In this article, we propose a novel federated unsupervised learning method for image classification without the use of any ground truth annotations. In IoT scenarios, a big challenge is that decentralized data among multiple clients is normally nonindependent and identically distributed (non-IID), leading to performance degradation. To address this issue, we further propose a dynamic update mechanism that can decide how to update the local model based on weights divergence. Extensive experiments show that our method outperforms all baseline methods by large margins, including +6.67% on CIFAR-10, +5.15% on STL-10, and +8.44% on SVHN in terms of classification accuracy. In particular, we obtain promising results on Mini-ImageNet and COVID-19 data sets and outperform several federated unsupervised learning methods under non-IID settings.

In this paper, we propose a method to address the issue of link loss inference in uncertain networks. Although numerous loss inference methods have been proposed in recent years, most of them ignore the unstable states of networks. That is, the performances of real network environments, such as link loss rates and end-to-end routes, are constantly changing. Ignoring these uncertain factors of the underlying network significantly hinders development of a solution. To address this problem, we propose a method to infer the link loss rates, even when the network is uncertain. After obtaining the routing matrix corresponding to the given topology, optimal probing paths are selected from all available paths to measure the end-to-end loss rates. According to the measurement results, each link is divided into different loss levels. Finally, we compute the loss range of each congested link by sample fitting. Compared with a state-of-the art method applied to realistic Internet service provider topologies, our algorithm not only required fewer injected probes, but it also increased the accuracy by 25 to 35%. The promising results demonstrate that our new method can be well applied to the practical uncertain networks.