ABSTRACT The sixth‐generation (6G) mobile network is expected to trigger an unprecedented surge in data traffic. unmanned aerial vehicle (UAV)‐assisted mobile edge computing (MEC) has emerged as a promising paradigm for emergency communications, yet dynamic user mobility and heterogeneous task demands pose severe challenges. Existing reinforcement learning (RL) methods (such as proximal policy optimization (PPO)), while effective in continuous control, often suffer from unstable convergence and neglect task‐priority differentiation, leading to excessive delays for critical tasks. To address these issues, we propose a novel KM‐PPO framework that integrates interference‐aware k‐means clustering with PPO. The clustering stage aggregates users by spatial distribution and task urgency, yielding an interference‐suppressed initial UAV deployment. The trajectory optimization stage employs PPO with clipped probability ratios, which serve to constrain policy updates and maintain stability, and a priority‐sensitive reward function, enabling UAVs to adaptively adjust motion trajectories under dynamic conditions. Compared with the baseline algorithms of k‐means integrated with twin delayed deep deterministic policy gradient (KM‐TD3) and k‐means integrated with soft actor‐critic (KM‐SAC), our method reduces average system latency by 15.9% and 24.6%, respectively, while achieving an 86.7% success rate for high‐priority tasks. These results demonstrate that KM‐PPO not only ensures stable convergence in the environments but also guarantees quality of service (QoS) for mission‐critical tasks, highlighting viability for UAV MEC deployments.

Progressive deep image compression (DIC) with hybrid contexts is an under-investigated problem that aims to jointly maximize the utility of a compressed image for multiple contexts or tasks under variable rates. In this paper, we consider the contexts of image reconstruction and classification. We propose a DIC framework, called residual-enhanced mask-based progressive generative coding (RMPGC), designed for explicit control of the performance within the rate-distortion-classification-perception (RDCP) trade-off. Three independent mechanisms are introduced to yield a semantically structured latent representation that can support parameterized control of rate and context adaptation. Experimental results show that the proposed RMPGC outperforms a benchmark DIC scheme using the same generative adversarial nets (GANs) backbone in all six metrics related to classification, distortion, and perception. Moreover, RMPGC is a flexible framework that can be applied to different neural network backbones. Some typical implementations are given and shown to outperform the classic BPG codec and four state-of-the-art DIC schemes in classification and perception metrics, with a slight degradation in distortion metrics. Our proposal of a nonlinear-neural-coded and richly structured latent space makes the proposed DIC scheme well suited for image compression in wireless communications, multi-user broadcasting, and multi-tasking applications.