ABSTRACT This study proposes a federated split learning framework for large language models (FedsLLM) integrated with rate‐splitting multiple access (RSMA), aimed at enhancing the efficiency and privacy of LLM training in wireless communication systems. By leveraging low‐rank adaptation (LoRA) to distribute computational loads and a fluid antenna system to dynamically optimize channel capacity, the framework effectively reduces training latency through joint optimization of learning accuracy and communication resources. Experimental results demonstrate that the proposed framework significantly outperforms traditional time‐division multiple access including time division multiple access, frequency division multiple access (FDMA), enhanced bandwidth FDMA, and fairness‐enhanced FDMA across multiple metrics: at a transmit power of 20 dBm, RSMA reduces task completion time by 8.3%; under 20 MHz bandwidth, it achieves a 25% performance improvement; and even with a data volume of 900 Kbits, it maintains a 12% advantage. The adopted alternating optimization algorithm converges rapidly, reaching 95% of the optimal value within only 5 iterations, substantially outperforming the fixed‐point method. Overall, FedsLLM‐RSMA effectively addresses privacy, computational and communication bottlenecks in distributed LLM training. Compared to TDMA, it reduces total training latency by 28% and improves communication efficiency by 35%, while achieving higher model accuracy and faster convergence. This work provides a viable pathway for efficient and scalable deployment of LLMs in 6G networks.

Uncrewed Aerial Vehicles (UAVs) have become key enabling platforms for low-altitude economic networks, yet achieving efficient and adaptive optimization under resource-constrained and dynamic environments remains challenging. This paper investigates language models for UAV-enabled Wireless Power Transfer (WPT) systems. First, a lightweight small language model (SLM)-based solution is developed using a pre-trained BERT backbone, enhanced UAV embeddings and contextual features, a geometry-aware path decoder, and ensemble inference to achieve low complexity, low latency, and high energy efficiency. Second, an Agentic AI-based framework is designed to exploit the reasoning and interactive capabilities of large language models (LLMs). It integrates four collaborative agents—Initializer, Actor, Critic, and Reflector—to form a closed loop of generation, optimization, evaluation, and reflection for iterative UAV path and energy optimization. Finally, simulations compare the SLM-, LLM-, and Agentic AI-based approaches.

With the advent of 6G communications, intelligent communication systems face multiple challenges, including constrained perception and response capabilities, limited scalability, and low adaptability in dynamic environments. To address these challenges, this tutorial provides a systematic and comprehensive introduction to the principles, design, and applications of Large Artificial Intelligence Models (LAMs) and Agentic AI technologies in intelligent communication systems, aiming to offer researchers an integrated overview of cutting-edge methodologies and practical insights. First, the tutorial outlines the background of 6G communications and reviews the technological evolution from LAMs to Agentic AI. It then systematically examines the key components required for constructing LAMs, classifies various types of LAMs, and analyzes their applicability in communication. A LAM-centric design paradigm tailored for communication systems is subsequently proposed, encompassing dataset construction, internal learning, and external learning approaches. Building upon this foundation, the tutorial develops an LAM-based Agentic AI system for intelligent communications, elaborating on its core components—including agents, world models, planners, knowledge bases, tools, and memory modules— as well as their interaction mechanisms. Finally, it provides an in-depth review of representative applications of LAMs and Agentic AI in communication scenarios, and summarizes the current research challenges and future directions, with the goal of fostering the development of efficient, secure, and sustainable next-generation intelligent communication systems.

Traditional unimodal sensing faces limitations in accuracy and capability, and its decoupled implementation with communication systems increases latency in bandwidth-constrained environments. Additionally, single-task-oriented sensing systems fail to address users’ diverse demands. To overcome these challenges, we propose a semantic-driven integrated multimodal sensing and communication (SIMAC) framework. This framework leverages a joint source-channel coding architecture to achieve simultaneous sensing, decoding, and transmission of sensing results. Specifically, SIMAC first introduces a multimodal semantic fusion (MSF) network, which employs two extractors to extract semantic information from radar signals and images, respectively. MSF then applies cross-attention mechanisms to fuse these unimodal features and generate multimodal semantic representations. Secondly, we present a large language model (LLM)-based semantic encoder (LSE), where relevant communication parameters and multimodal semantics are mapped into a unified latent space and input to the LLM, enabling channel-adaptive semantic encoding. Thirdly, a task-oriented sensing semantic decoder (SSD) is proposed, in which different decoded heads are designed according to the specific needs of tasks. Simultaneously, a multi-task learning strategy is introduced to train the SIMAC framework, achieving diverse sensing services. Finally, experimental simulations demonstrate that the proposed framework achieves diverse and higher-accuracy sensing services.

The rapid development of generative Artificial Intelligence (AI) continually unveils the potential of Semantic Communication (SemCom). However, current talking-face SemCom systems still encounter challenges such as low bandwidth utilization, semantic ambiguity, and diminished Quality of Experience (QoE). This study introduces a Large Generative Model-assisted Talking-face Semantic Communication (LGM-TSC) System tailored for talking-face video communication. Firstly, we introduce a Generative Semantic Extractor (GSE) at the transmitter based on the FunASR model to convert semantically sparse talking-face videos into text with high information density. Secondly, we establish a private Knowledge Base (KB) based on the Large Language Model (LLM) for semantic disambiguation and correction, complemented by a joint knowledge base-semantic-channel coding scheme. Finally, at the receiver, we propose a Generative Semantic Reconstructor (GSR) that utilizes BERT-VITS2 and SadTalker models to transform text back into a high-QoE talking-face video matching the user’s timbre. Simulation results demonstrate the feasibility and effectiveness of the proposed LGM-TSC system.

The intelligent reflecting surface (IRS) and unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) system is widely used in temporary and emergency scenarios. Our goal is to minimize the energy consumption of the MEC system by jointly optimizing UAV locations, IRS phase shift, task offloading, and resource allocation with a variable number of UAVs. To this end, we propose a flexible resource scheduling (FRES) framework by employing a novel deep progressive reinforcement learning that includes the following innovations. First, a novel multitask agent is presented to deal with the mixed integer nonlinear programming (MINLP) problem. The multitask agent has two output heads designed for different tasks, in which a classified head is employed to make offloading decisions with integer variables while a fitting head is applied to solve resource allocation with continuous variables. Second, a progressive scheduler is introduced to adapt the agent to the varying number of UAVs by progressively adjusting a part of neurons in the agent. This structure can naturally accumulate experiences and be immune to catastrophic forgetting. Finally, a light taboo search (LTS) is introduced to enhance the global search of the FRES. The numerical results demonstrate the superiority of the FRES framework, which can make real-time and optimal resource scheduling even in dynamic MEC systems.

Semantic Communication (SC) has emerged as a novel communication paradigm in recent years. Nevertheless, extant Image Semantic Communication (ISC) systems face several challenges in dynamic environments, including low information density, catastrophic forgetting, and uncertain Signal-to-Noise Ratio (SNR). To address these challenges, we propose a novel Vision-Language Model-based Cross-modal Semantic Communication (VLM-CSC) system. The VLM-CSC comprises three novel components: 1) Cross-modal Knowledge Base (CKB) is used to extract high-density textual semantics from the semantically sparse image at the transmitter and reconstruct the original image based on textual semantics at the receiver. The transmission of high-density semantics contributes to alleviating bandwidth pressure; 2) Memory-assisted Encoder and Decoder (MED) employ a hybrid long/short-term memory mechanism, enabling the semantic encoder and decoder to overcome catastrophic forgetting in dynamic environments when there is a drift in the distribution of semantic features; 3) Noise Attention Module (NAM) employs attention mechanisms to adaptively adjust the semantic coding and the channel coding based on SNR, ensuring the robustness of the CSC system. The experimental simulations validate the effectiveness, adaptability, and robustness of the CSC system.

In fire surveillance, Industrial Internet of Things (IIoT) devices require transmitting large monitoring data frequently, which leads to huge consumption of spectrum resources. Hence, we propose an Industrial Edge Semantic Network to allow IIoT devices to send warnings through Semantic communication (SC). Thus, we should consider 1) data privacy and security; 2) SC model adaptation for heterogeneous devices; 3) explainability of semantics. Therefore, first, we present an eXplainable Semantic Federated Learning (XSFL) to train the SC model, thus ensuring data privacy and security. Then, we present an adaptive client training strategy to provide a specific SC model for each device according to its Fisher information matrix, thus overcoming the heterogeneity. Next, an Explainable SC mechanism is designed, which introduces a leakyReLU-based activation mapping to explain the relationship between the extracted semantics and monitoring data. Finally, simulation results demonstrate the effectiveness of XSFL.

This article studies a dynamic Mobile Edge Computing (MEC) system assisted by Unmanned Aerial Vehicles (UAVs) and Intelligent Reflective Surfaces (IRSs). We propose a scaleable resource scheduling algorithm to minimize the energy consumption of all UEs and UAVs in the MEC system with a variable number of UAVs. We propose a Multi-tAsk Resource Scheduling (MARS) framework based on Deep Reinforcement Learning (DRL) to solve the problem. First, we present a novel Advantage Actor-Critic (A2C) structure with the state-value critic and entropy-enhanced actor to reduce variance and enhance the policy search of DRL. Then, we present a multi-head agent with three different heads in which a classification head is applied to make offloading decisions and a regression head is presented to allocate computational resources, and a critic head is introduced to estimate the state value of the selected action. Next, we introduce a multi-task controller to adjust the agent to adapt to the varying number of UAVs by loading or unloading a part of weights in the agent. Finally, a Light Wolf Search (LWS) is introduced as the action refinement to enhance the exploration in the dynamic action space. The numerical results demonstrate the feasibility and efficiency of the MARS framework.

Unmanned aerial vehicles (UAVs) can be applied in many Internet of Things (IoT) systems, e.g., smart farms, as a data collection platform. However, the UAV-IoT wireless channels may be occasionally blocked by trees or high-rise buildings. An intelligent reflecting surface (IRS) can be applied to improve the wireless channel quality by smartly reflecting the signal via a large number of low-cost passive reflective elements. This article aims to minimize the energy consumption of the system by jointly optimizing the deployment and trajectory of the UAV. The problem is formulated as a mixed-integer-and-nonlinear programming (MINLP), which is challenging to address by the traditional solution, because the solution may easily fall into the local optimal. To address this issue, we propose a joint optimization framework of deployment and trajectory (JOLT), where an adaptive whale optimization algorithm (AWOA) is applied to optimize the deployment of the UAV, and an elastic ring self-organizing map (ERSOM) is introduced to optimize the trajectory of the UAV. Specifically, in AWOA, a variable-length population strategy is applied to find the optimal number of stop points, and a nonlinear parameter$a$and a partial mutation rule are introduced to balance the exploration and exploitation. In ERSOM, a competitive neural network is also introduced to learn the trajectory of the UAV by competitive learning, and a ring structure is presented to avoid the trajectory intersection. Extensive experiments are carried out to show the effectiveness of the proposed JOLT framework.

In large-scale mobile edge computing (MEC) systems, the task latency, and energy consumption are important for massive resource-consuming and delay-sensitive Internet of Things Devices (IoTDs). Against this background, we propose a distributed intelligent resource scheduling (DIRS) framework to minimize the sum of task latency and energy consumption for all IoTDs, which can be formulated as a mixed-integer nonlinear programming. The DIRS framework includes centralized training relying on the global information and distributed decision making by each agent deployed in each MEC server. Specifically, we first introduce a novel multiagent ensemble-assisted distributed deep reinforcement learning (DRL) architecture, which can simplify the overall neural network structure of each agent by partitioning the state space and also improve the performance of a single agent by combining decisions of all the agents. Second, we apply action refinement to enhance the exploration ability of the proposed DIRS framework, where the near-optimal state-action pairs are obtained by a novel Levy flight search. Finally, an imitation acceleration scheme is presented to pretrain all the agents, which can significantly accelerate the learning process of the proposed framework through learning the professional experience from a small amount of demonstration data. The simulation results in three typical scenarios demonstrate that the proposed DIRS framework is efficient and outperforms the existing benchmark schemes.

This study aims to develop a novel wavelet neural-network (WNN) model for solving electrical resistivity imaging (ERI) inversion with massive amounts of measured data in control and measurement fields. In the proposed method, we design a mixed multilayer WNN (MMWNN) which uses Morlet and Mexican wavelons as different activation functions in a cascaded hidden layer structure. Meanwhile, a hybrid STGWO-GD learning approach is used to improve the learning ability of the MMWNN, which is a combination of the self-tuning grey wolf optimizer (STGWO) and the gradient descent (GD) algorithm adopting the advantages of each other. Moreover, updating formulas of the GD algorithm are derived, and a Gaussian updating operator with weighted hierarchical hunting, a chaotic oscillation equation, and a nonlinear modulation coefficient are introduced to improve the hierarchical hunting and the control parameter adjustment of the modified STGWO. Five examples are used with the aim of assessing the availability and feasibility of the proposed inversion method. The inversion results are promising and show that the introduced method is superior to other competitors in terms of inversion accuracy and computational efficiency. Furthermore, the effectiveness of the proposed method is demonstrated over a classical benchmark successfully.

An online resource scheduling framework is proposed for minimizing the sum of weighted task latency for all the Internet-of-Things (IoT) users, by optimizing offloading decision, transmission power, and resource allocation in the large-scale mobile-edge computing (MEC) system. Toward this end, a deep reinforcement learning (DRL)-based solution is proposed, which includes the following components. First, a related and regularized stacked autoencoder (2r-SAE) with unsupervised learning is applied to perform data compression and representation for high-dimensional channel quality information (CQI) data, which can reduce the state space for DRL. Second, we present an adaptive simulated annealing approach (ASA) as the action search method of DRL, in which an adaptive ${h}$ -mutation is used to guide the search direction and an adaptive iteration is proposed to enhance the search efficiency during the DRL process. Third, a preserved and prioritized experience replay (2p-ER) is introduced to assist the DRL to train the policy network and find the optimal offloading policy. The numerical results are provided to demonstrate that the proposed algorithm can achieve near-optimal performance while significantly decreasing the computational time compared with existing benchmarks.

In this article, we consider a hybrid mobile edge computing (H-MEC) platform, which includes ground stations (GSs), ground vehicles (GVs), and unmanned aerial vehicles (UAVs), all with the mobile edge cloud installed to enable user equipments (UEs) or Internet of Things (IoT) devices with intensive computing tasks to offload. Our objective is to obtain an online offloading algorithm to minimize the energy consumption of all the UEs, by jointly optimizing the positions of GVs and UAVs, user association and resource allocation in real time, while considering the dynamic environment. To this end, we propose a hybrid deep-learning-based online offloading (H2O) framework where a large-scale path-loss fuzzy c-means (LS-FCM) algorithm is first proposed and used to predict the optimal positions of GVs and UAVs. Second, a fuzzy membership matrix U-based particle swarm optimization (U-PSO) algorithm is applied to solve the mixed-integer nonlinear programming (MINLP) problems and generate the sample data sets for the deep neural network (DNN) where the fuzzy membership matrix can capture the small-scale fading effects and the information of mutual interference. Third, a DNN with the scheduling layer is introduced to provide the user association and computing resource allocation under the practical latency requirement of the tasks and limited available computing resource of H-MEC. In addition, different from the traditional DNN predictor, we only input one UE's information to the DNN at one time, which will be suitable for the scenarios where the number of UE is varying and avoid the curse of dimensionality in DNN.