The rapid growth of web content has spurred the widespread adoption of on-device AI assistants powered by large language models (LLMs). However, deploying and personalizing these assistants in real-world environments remains challenging due to limited annotation budgets and scarce on-device fine-tuning resources. Existing edge–cloud collaboration frameworks typically rely on costly cloud-based supervision or perform full-layer finetuning, leading to inefficiencies in both computation and adaptation. To address these limitations, we propose EcoTune, a budget-constrained framework for efficient edge–cloud collaborative adaptation. EcoTune jointly optimizes representative data selection for cloud annotation and selective on-device model adaptation within a unified closed-loop process. Specifically, it employs a multi-armed bandit–based strategy to identify highvalue user interactions for cloud supervision and a layer importance–driven adaptation mechanism to update only critical components of the small language model (SLM). This coordinated optimization enables dynamic, resource-efficient personalization under stringent annotation and tuning budgets. Experiments on real-world testbeds demonstrate that EcoTune achieves up to 20%-60% reduction in annotation costs and significantly lowers fine-tuning memory consumption compared to state-of-the-art baselines, providing a practical and scalable solution for personalized on-device LLMs.

Deep neural networks (DNNs) such as convolutional networks and Transformers are increasingly deployed to provide intelligent service. However, enabling large-scale DNNs in an infrastructure-less mobile crowd encounters the challenges of high resource consumption, poor performance, and low service availability. This paper proposes CrowdNet, a novel device-todevice (D2D) collaborative inference framework for dynamic mobile environments. CrowdNet introduces a CellNet architecture as its core component, designed as lightweight DNNs ideal for deployment and operation on resource-constrained mobile devices. The CellNets can perform inference tasks either independently or collaboratively to ensure robustness against network disruptions. A topology expansion method is utilized to create an inference flow from the physical communication topology, enabling the distributed operation of inference tasks. To handle the dynamic participation of mobile devices, CrowdNet employs fine-tuning adaptation for flexible assembly and collaborative inference. A reinforcement learning (RL)-based approach is introduced to optimize inference topology. Trained with a multiobjective optimization strategy, CrowdNet can enhance overall performance while maintaining individual CellNet functionality. Extensive experiments based on mobile network testbed and realworld datasets validate the effectiveness of CrowdNet on various intelligent tasks, exhibiting remarkable performance gains and robustness compared to state-of-the-art approaches.

Unmanned Aerial Vehicles (UAVs) can be used as aerial base stations for data collection in next-generation wireless networks due to their high adaptability and maneuverability. This paper investigates the scenario where multiple UAVs cooperatively fly over heterogeneous ground users (GUs) and collect data without a central controller. With the consideration of signal-to-interference-and-noise ratio (SINR) and fairness among users, we jointly optimize the trajectories of UAVs and the GUs associations to maximize the total throughput and energy efficiency. We formulate the long-term optimization problem as a decentralized partially observed Markov decision processes (DEC-POMDP) and derive an approach combining the coalition formation game (CFG) and multi-agent deep reinforcement learning (MADRL). We first formulate the discrete association scheduling problem as a non-cooperative theoretical game and use the CFG algorithm to achieve a decentralized scheme converging to Nash equilibrium (NE). Then, a MARL-based technique is developed to optimize the trajectories and energy consumption continuously in a centralized-training but decentralized-execution manner. Simulation results demonstrate that the proposed algorithm outperforms the commonly used schemes in the literature, regarding the fair throughput and energy consumption in a distributed manner.