Distributed Artificial Intelligence (AI) model training over mobile edge networks encounters significant challenges due to the data and resource heterogeneity of edge devices. The former hampers the convergence rate of the global model, while the latter diminishes the devices' resource utilization efficiency. In this paper, we propose a generative AI-empowered federated learning to address these challenges by leveraging the idea of FIlling the MIssing (FIMI) portion of local data. Specifically, FIMI can be considered as a resource-aware data augmentation method that effectively mitigates the data heterogeneity while ensuring efficient FL training. We first quantify the relationship between the training data amount and the learning performance. We then study the FIMI optimization problem with the objective of minimizing the device-side overall energy consumption subject to required learning performance constraints. The decomposition-based analysis and the cross-entropy searching method are leveraged to derive the solution, where each device is assigned suitable AI-synthesized data and resource utilization policy. Experiment results demonstrate that FIMI can save up to 50% of the device-side energy to achieve the target global test accuracy in comparison with the existing methods. Meanwhile, FIMI can significantly enhance the converged global accuracy under the non-independently-and-identically distribution (non-IID) data.
An LSTM-aided hybrid random access scheme (LSTMH-RA) is proposed to support diverse quality of service (QoS) requirements in 6G MTC heterogeneous networks where URLLC and mMTC devices coexist. This scheme employs an attention-based LSTM prediction model to predict the number of active URLLC devices, determines the parameters of the multi-user detection algorithm dynamically, and then allows URLLC devices to access the network via a two-step contention-free access procedure, to meet latency and reliability access requirements; mMTC devices access the network via a contentionbased TA-aided access mechanism to meet massive access requirement. We analyze the successful access probability of the LSTMH-RA scheme. Numerical results show that, compared to the benchmark schemes, the LSTMH-RA scheme significantly improves the successful access probability, and satisfies the diverse QoS requirements of URLLC and mMTC devices