By provisioning inference offloading services, edge inference drives the rapid growth of AI applications at the network edge. However, achieving high task throughput with stringent latency requirements remains a significant challenge. To address this issue, we develop a parameter-sharing AI model loading (PartialLoading) framework for multi-user edge inference, which exploits two key insights: 1) the majority of latency arises from loading AI models into server GPU memory, and 2) different AI models can share a significant number of parameters, for which redundant loading should be avoided. Towards this end, we formulate a joint multi-user scheduling and spectrum bandwidth allocation problem to maximize task throughput by exploiting shared parameter blocks across models. The intuition is to judiciously schedule user requests to reuse the shared parameter blocks between consecutively loaded models, thereby reducing model loading time substantially. To facilitate solution finding, we decouple the problem into two sub-problems, i.e., user scheduling and bandwidth allocation, showing that solving them sequentially is equivalent to solving the original problem. Due to the NP-hardness of the problem, we first study an important special case called the "bottom-layer-sharing" case, where AI models share some bottom layers within clusters, and design a dynamic programming-based algorithm to obtain the optimal solution in polynomial time. For the general case, where shared parameter blocks appear at arbitrary positions within AI models, we propose a greedy heuristic to obtain the sub-optimal solution efficiently. Simulation results demonstrate that the proposed framework significantly improves task throughput under deadline constraints compared with user scheduling without exploiting parameter sharing.

The rise of generative artificial intelligence (AI) as a novel frontier that uniquely merges advanced levels of intelligence with revolutionary user experiences is redefining the AI landscape for future cellular networks. In particular, the transition towards 6G systems has introduced a myriad of challenges inherent to their AI-native network design, requiring innovative solutions to enable real-time network orchestration, intelligent decision-making, and adaptive dynamic configurations. Meanwhile, the envisioned user experiences for 6G are growing increasingly complex, exceeding the capabilities offered by vintage wireless technologies and conventional AI solutions to satisfy their advanced demands. With its disruptive impact evident across diverse fields, generative AI possesses immense potential to tackle these challenges, leveraging its exceptional capabilities to manage complex tasks, operate autonomously, and adapt seamlessly to scenarios beyond its training domain. Remarkably, generative AI provides a transformative opportunity for telecom and cellular networks to bridge this defined gap in 6G systems, thereby shifting towards a new era with cutting-edge AI innovations across the different system and user levels.

Recent breakthroughs in artificial intelligence (AI), wireless communications, and sensing technologies have accelerated the evolution of edge intelligence. However, conventional systems still grapple with issues such as low communication efficiency, redundant data acquisition, and poor model generalization. To overcome these challenges, we propose an innovative framework that enhances edge intelligence through AI-in-the-loop joint sensing and communication (JSAC). This framework features an AI-driven closed-loop control architecture that jointly optimizes system resources, thereby delivering superior system-level performance. A key contribution of our work is establishing an explicit relationship between validation loss and the system's tunable parameters. This insight enables dynamic reduction of the generalization error through AI-driven closed-loop control. Specifically, for sensing control, we introduce an adaptive data collection strategy based on gradient importance sampling, allowing edge devices to autonomously decide when to terminate data acquisition and how to allocate sample weights based on real-time model feedback. For communication control, drawing inspiration from stochastic gradient Langevin dynamics (SGLD), our joint optimization of transmission power and batch size converts channel and data noise into gradient perturbations that help mitigate overfitting. Experimental evaluations demonstrate that our framework reduces communication energy consumption by up to 77 percent and sensing costs measured by the number of collected samples by up to 52 percent while significantly improving model generalization -- with up to 58 percent reductions of the final validation loss. It validates that the proposed scheme can harvest the mutual benefit of AI and JSAC systems by incorporating the model itself into the control loop of the system.

Rare events, despite their infrequency, often carry critical information and require immediate attentions in mission-critical applications such as autonomous driving, healthcare, and industrial automation. The data-intensive nature of these tasks and their need for prompt responses, combined with designing edge AI (or edge inference), pose significant challenges in systems and techniques. Existing edge inference approaches often suffer from communication bottlenecks due to high-dimensional data transmission and fail to provide timely responses to rare events, limiting their effectiveness for mission-critical applications in the sixth-generation (6G) mobile networks. To overcome these challenges, we propose a channel-adaptive, event-triggered edge-inference framework that prioritizes efficient rare-event processing. Central to this framework is a dual-threshold, multi-exit architecture, which enables early local inference for rare events detected locally while offloading more complex rare events to edge servers for detailed classification. To further enhance the system's performance, we developed a channel-adaptive offloading policy paired with an online algorithm to dynamically determine the optimal confidence thresholds for controlling offloading decisions. The associated optimization problem is solved by reformulating the original non-convex function into an equivalent strongly convex one. Using deep neural network classifiers and real medical datasets, our experiments demonstrate that the proposed framework not only achieves superior rare-event classification accuracy, but also effectively reduces communication overhead, as opposed to existing edge-inference approaches.

Federated Dropout is an efficient technique to overcome both communication and computation bottlenecks for deploying federated learning at the network edge. In each training round, an edge device only needs to update and transmit a sub-model, which is generated by the typical method of dropout in deep learning, and thus effectively reduces the per-round latency. \textcolor{blue}{However, the theoretical convergence analysis for Federated Dropout is still lacking in the literature, particularly regarding the quantitative influence of dropout rate on convergence}. To address this issue, by using the Taylor expansion method, we mathematically show that the gradient variance increases with a scaling factor of $\gamma/(1-\gamma)$, with $\gamma \in [0, \theta)$ denoting the dropout rate and $\theta$ being the maximum dropout rate ensuring the loss function reduction. Based on the above approximation, we provide the convergence analysis for Federated Dropout. Specifically, it is shown that a larger dropout rate of each device leads to a slower convergence rate. This provides a theoretical foundation for reducing the convergence latency by making a tradeoff between the per-round latency and the overall rounds till convergence. Moreover, a low-complexity algorithm is proposed to jointly optimize the dropout rate and the bandwidth allocation for minimizing the loss function in all rounds under a given per-round latency and limited network resources. Finally, numerical results are provided to verify the effectiveness of the proposed algorithm.

To bridge the digital divide, the space-ground integrated networks (SGINs), which will be a key component of the six-generation (6G) mobile networks, are expected to deliver artificial intelligence (AI) services to every corner of the world. One mission of SGINs is to support federated learning (FL) at a global scale. However, existing space-ground integrated FL frameworks involve ground stations or costly inter-satellite links, entailing excessive training latency and communication costs. To overcome these limitations, we propose an infrastructure-free federated learning framework based on a model dispersal (FedMeld) strategy, which exploits periodic movement patterns and store-carry-forward capabilities of satellites to enable parameter mixing across large-scale geographical regions. We theoretically show that FedMeld leads to global model convergence and quantify the effects of round interval and mixing ratio between adjacent areas on its learning performance. Based on the theoretical results, we formulate a joint optimization problem to design the staleness control and mixing ratio (SC-MR) for minimizing the training loss. By decomposing the problem into sequential SC and MR subproblems without compromising the optimality, we derive the round interval solution in a closed form and the mixing ratio in a semi-closed form to achieve the \textit{optimal} latency-accuracy tradeoff. Experiments using various datasets demonstrate that FedMeld achieves superior model accuracy while significantly reducing communication costs as compared with traditional FL schemes for SGINs.

LoLaFL: Low-Latency Federated Learning via Forward-only Propagation Jierui Zhang, Graduate Student Member, IEEE, Jianhao Huang, Member, IEEE, and Kaibin Huang, Fellow, IEEE Abstract --Federated learning (FL) has emerged as a widely adopted paradigm for enabling edge learning with distributed data while ensuring data privacy. However, the traditional FL with deep neural networks trained via backpropagation can hardly meet the low-latency learning requirements in the sixth generation (6G) mobile networks. This challenge mainly arises from the high-dimensional model parameters to be transmitted and the numerous rounds of communication required for convergence due to the inherent randomness of the training process. T o address this issue, we adopt the state-of-the-art principle of maximal coding rate reduction to learn linear discriminative features and extend the resultant white-box neural network into FL, yielding the novel framework of Low-Latency Federated Learning (LoLaFL) via forward-only propagation. LoLaFL enables layer-wise transmissions and aggregation with significantly fewer communication rounds, thereby considerably reducing latency. Additionally, we propose two nonlinear aggregation schemes for LoLaFL. The first scheme is based on the proof that the optimal NN parameter aggregation in LoLaFL should be harmonic-mean-like. The second scheme further exploits the low-rank structures of the features and transmits the low-rank-approximated covariance matrices of features to achieve additional latency reduction. Theoretic analysis and experiments are conducted to evaluate the performance of LoLaFL. In comparison with traditional FL, the two nonlinear aggregation schemes for LoLaFL can achieve reductions in latency of over 91% and 98%, respectively, while maintaining comparable accuracies. I NTRODUCTION With the growing volume of data and the increasing number of edge devices, the sixth generation (6G) mobile networks are envisioned to support a wide range of AI-based applications at the network edge, including augmented/mixed/virtual reality, connected robotics and autonomous systems, and smart cities and homes, among others [1], [2]. To realize this vision, researchers have been motivated to develop technologies to deploy AI models at the network edge. These technologies, collectively called edge learning, leverage the mobile-edge-computing platform to train edge-AI models among edge servers and devices [3], [4]. For its preservation of data privacy, federated learning (FL) emerges as a widely adopted solution for distributed edge learning, where local models are trained using local devices' data and sent to the server for updating the global model [5]-[8]. This collaborative training approach enables multiple devices and a server to train a global J. Zhang, J. Huang, and K. Huang are with the Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong. However, FL faces its own challenges.

Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, leading to their increasing deployment in wireless networks for a wide variety of user services. However, the growing longer prompt setting highlights the crucial issue of computational resource demands and huge communication load. To address this challenge, we propose Joint Power and Prompt Optimization (JPPO), a framework that combines Small Language Model (SLM)-based prompt compression with wireless power allocation optimization. By deploying SLM at user devices for prompt compression and employing Deep Reinforcement Learning for joint optimization of compression ratio and transmission power, JPPO effectively balances service quality with resource efficiency. Experimental results demonstrate that our framework achieves high service fidelity and low bit error rates while optimizing power usage in wireless LLM services. The system reduces response time by about 17%, with the improvement varying based on the length of the original prompt.

The emergence of large-scale foundation models (FoMo's) that can perform human-like intelligence motivates their deployment at the network edge for devices to access state-of-the-art artificial intelligence. For better user experiences, the pre-trained FoMo's need to be adapted to specialized downstream tasks through fine-tuning techniques. To transcend a single device's memory and computation limitations, we advocate multi-device cooperation within the device-edge cooperative fine-tuning (DEFT) paradigm, where edge devices cooperate to simultaneously optimize different parts of fine-tuning parameters within a FoMo. However, the parameter blocks reside at different depths within a FoMo architecture, leading to varied computation latency-and-memory cost due to gradient backpropagation-based calculations. The heterogeneous on-device computation and memory capacities and channel conditions necessitate an integrated communication-and-computation allocation of local computation loads and communication resources to achieve low-latency (LoLa) DEFT. To this end, we consider the depth-ware DEFT block allocation problem. The involved optimal block-device matching is tackled by the proposed low-complexity Cutting-RecoUNting-CHecking (CRUNCH) algorithm, which is designed by exploiting the monotone-increasing property between block depth and computation latency-and-memory cost. Next, the joint bandwidth-and-block allocation makes the problem more sophisticated. We observe a splittable Lagrangian expression through the transformation and analysis of the original problem, where the variables indicating device involvement are introduced. Then, the dual ascent method is employed to tackle this problem iteratively. Through extensive experiments conducted on the GLUE benchmark, our results demonstrate significant latency reduction achievable by LoLa DEFT for fine-tuning a RoBERTa model.

Next-generation mobile networks are expected to facilitate fast AI model downloading to end users. By caching models on edge servers, mobile networks can deliver models to end users with low latency, resulting in a paradigm called edge model caching. In this paper, we develop a novel model placement scheme, called parameter-sharing model caching (TrimCaching). TrimCaching exploits the key observation that a wide range of AI models, such as convolutional neural networks or large language models, can share a significant proportion of parameter blocks containing reusable knowledge, thereby improving storage efficiency. To this end, we formulate a parameter-sharing model placement problem to maximize the cache hit ratio in multi-edge wireless networks by balancing the fundamental tradeoff between storage efficiency and service latency. We show that the formulated problem is a submodular maximization problem with submodular constraints, for which no polynomial-time approximation algorithm exists. To overcome this challenge, we study an important special case, where a small fixed number of parameter blocks are shared across models, which often holds in practice. In such a case, a polynomial-time algorithm with $\left(1-\epsilon\right)/2$-approximation guarantee is developed. Subsequently, we address the original problem for the general case by developing a greedy algorithm. Simulation results demonstrate that the proposed TrimCaching framework significantly improves the cache hit ratio compared with state-of-the-art content caching without exploiting shared parameters in AI models.