Traffic flow estimation (TFE) is crucial for intelligent transportation systems, which traditionally rely on extensive road sensors at significant cost. Sparse mobile crowdsensing (MCS) enables a cost-effective alternative by utilizing sparsely distributed probe vehicle data (PVD) collected from connected vehicles. However, according to the central limit theorem, the initial TFE results derived from sparse PVD by calculating the average vehicle speed for each road segment turn out to be inaccurate. To this end, this paper proposes a novel sparse MCS TFE approach and improves its accuracy with the spatial-temporal generative artificial intelligence (GAI) framework. Within this framework, the conditional encoder mines spatial-temporal correlations in the initial TFE results, and the generative decoder generates accurate TFE outputs, achieving similar results to those employing complete and sufficient PVD. Additionally, we discuss the design of the spatial-temporal neural network that serves as the backbone of the conditional encoder for effectively capturing spatial-temporal correlations. The effectiveness of the proposed TFE approach is demonstrated through evaluations based on real-world connected vehicle data. Experimental results demonstrate the feasibility of the proposed TFE approach and highlight the effectiveness of the spatial-temporal GAI framework in improving estimation accuracy.

Semantic communication (SemCom) has emerged as a transformative paradigm for future wireless networks, aiming to improve communication efficiency by transmitting only the semantic meaning (or its encoded version) of the source data rather than the complete set of bits (symbols). However, traditional deep learning-based SemCom systems present challenges such as limited generalization, low robustness, and inadequate reasoning capabilities, primarily due to the inherently discriminative nature of deep neural networks. To address these limitations, generative artificial intelligence (GAI) is seen as a promising solution, offering notable advantages in learning complex data distributions, transforming data between high- and low-dimensional spaces, and generating high-quality content.This paper explores the applications of GAI in SemCom and presents a comprehensive study. It begins by introducing three widely used SemCom systems enabled by classical GAI models: variational autoencoders, generative adversarial networks, and diffusion models. For each system, the fundamental concept of the GAI model, the corresponding SemCom architecture, and a literature review of recent developments are provided. Subsequently, a novel generative SemCom system is proposed, incorporating cutting-edge GAI technology—large language models (LLMs). This system features LLM-based artificial intelligence (AI) agents at both the transmitter and receiver, which act as “brains” to enable advanced information understanding and content regeneration capabilities, respectively. Unlike traditional systems that focus on bitstream recovery, this design allows the receiver to directly generate the desired content from the coded semantic information sent by the transmitter. As a result, the communication paradigm shifts from “information recovery” to “information regeneration,” marking a new era in generative SemCom. A case study on point-to-point video retrieval is presented to demonstrate the effectiveness of the proposed system, showing a 99.98% reduction in communication overhead and a 53% improvement in average retrieval accuracy compared to traditional communication systems. Furthermore, four typical application scenarios for generative SemCom are described, followed by a discussion of three open issues for future research. In summary, this paper provides a comprehensive set of guidelines for applying GAI in SemCom, laying the groundwork for the efficient deployment of generative SemCom in future wireless networks.

This paper introduces a two-stage generative AI (GenAI) framework tailored for temporal spectrum cartography in low-altitude economy networks (LAENets). LAENets, characterized by diverse aerial devices such as UAVs, rely heavily on wireless communication technologies while facing challenges, including spectrum congestion and dynamic environmental interference. Traditional spectrum cartography methods have limitations in handling the temporal and spatial complexities inherent to these networks. Addressing these challenges, the proposed framework first employs a Reconstructive Masked Autoencoder (RecMAE) capable of accurately reconstructing spectrum maps from sparse and temporally varying sensor data using a novel dual-mask mechanism. This approach significantly enhances the precision of reconstructed radio frequency (RF) power maps. In the second stage, the Multi-agent Diffusion Policy (MADP) method integrates diffusion-based reinforcement learning to optimize the trajectories of dynamic UAV sensors. By leveraging temporal-attention encoding, this method effectively manages spatial exploration and exploitation to minimize cumulative reconstruction errors. Extensive numerical experiments show that this integrated GenAI framework consistently surpasses traditional interpolation and deep learning methods, especially under sparse sensing conditions. The proposed trajectory planner substantially improves spectrum map accuracy, reconstruction stability, and sensor deployment efficiency in dynamically evolving low-altitude environments.

Discrete representation has emerged as a powerful tool in task-oriented semantic communication (ToSC), offering compact, interpretable, and efficient representations well-suited for low-power edge intelligence scenarios. Its inherent digital nature aligns seamlessly with hardware-friendly deployment and robust storage/transmission protocols. However, current ToSC frameworks often decouple semantic-aware discrete mapping from the underlying channel characteristics and task demands, which leads to suboptimal communication performance and degraded task utility under variable wireless conditions. To address these limitations, this paper proposes a channel-aware discrete semantic coding framework tailored for low-power edge networks. Leveraging a Wasserstein-regularized objective, our approach (WS-VQ) aligns discrete code activations with optimal input distributions, thereby improving semantic fidelity, robustness, and task accuracy. Extensive experiments on the inference tasks across diverse signal-to-noise ratio (SNR) regimes show that our method achieves notable gains in accuracy and communication efficiency. This work provides new insights into integrating discrete semantics and channel optimization, paving the way for the widespread adoption of semantic communication in future digital infrastructures.

Enhancing future wireless networks presents a significant challenge for networking systems due to diverse user demands and the emergence of 6G technology. While reinforcement learning (RL) is a powerful framework, it often encounters difficulties with high-dimensional state spaces and complex environments, leading to substantial computational demands, distributed intelligence, and potentially inconsistent outcomes. Large language models (LLMs), with their extensive pretrained knowledge and advanced reasoning capabilities, offer promising tools to enhance RL in optimizing 6G wireless networks. We explore RL models augmented by LLMs, emphasizing their roles and the potential benefits of their synergy in wireless network optimization. We then examine LLM-enabled RL across various protocol layers: physical, data link, network, transport, and application layers. Additionally, we propose an LLM-assisted state representation and semantic extraction to enhance the multi-agent reinforcement learning (MARL) framework. This approach is applied to service migration and request routing, as well as topology graph generation in unmanned aerial vehicle (UAV)-satellite networks. Through case studies, we demonstrate that our framework effectively performs optimization of wireless network. Finally, we outline prospective research directions for LLM-enabled RL in wireless network optimization.

Semantic communication (SemCom) has emerged as a promising paradigm for achieving unprecedented communication efficiency in sixth-generation (6G) networks. However, deploying SemCom over digital systems introduces new challenges in maintaining robustness against packet losses that distort critical content. To address this issue, this paper proposes a novel framework, termed generative feature imputing, which comprises three key techniques. First, we introduce a spatial-error-concentration packetization strategy that spatially concentrates feature distortions by encoding feature elements based on their channel mappings. Second, building on this strategy, we propose a generative feature imputing method that utilizes a diffusion model to efficiently reconstruct missing features caused by packet losses. Finally, we develop a semantic-aware power allocation scheme that enables unequal error protection by allocating transmission power according to the semantic importance of each packet. Experimental results demonstrate that the proposed framework outperforms conventional approaches, such as Deep Joint Source-Channel Coding (DJSCC) and JPEG2000, under block fading conditions, achieving higher semantic accuracy and lower Learned Perceptual Image Patch Similarity (LPIPS) scores.

The growing demand for low-latency computing in 6G is driving the use of UAV-based low-altitude mobile edge computing (MEC) systems. However, limited spectrum often leads to severe uplink interference among ground terminals (GTs). In this paper, we investigate a rate-splitting multiple access (RSMA)-enabled low-altitude MEC system, where a UAV-based edge server assists multiple GTs in concurrently offloading their tasks over a shared uplink. We formulate a joint optimization problem involving the UAV 3D trajectory, RSMA decoding order, task offloading decisions, and resource allocation, aiming to mitigate multi-user interference and maximize energy efficiency. Given the high dimensionality, non-convex nature, and dynamic characteristics of this optimization problem, we propose a generative AI-enhanced deep reinforcement learning (DRL) framework to solve it efficiently. Specifically, we embed a diffusion model into the actor network to generate high-quality action samples, improving exploration in hybrid action spaces and avoiding local optima. In addition, a priority-based RSMA decoding strategy is designed to facilitate efficient successive interference cancellation with low complexity. Simulation results demonstrate that the proposed method for low-altitude MEC systems outperforms baseline methods, and that integrating GDM with RSMA can achieve significantly improved energy efficiency performance.

Digital task-oriented semantic communication (ToSC) aims to transmit only task-relevant information, significantly reducing communication overhead. Existing ToSC methods typically rely on learned codebooks to encode semantic features and map them to constellation symbols. However, these codebooks are often sparsely activated, resulting in low spectral efficiency and underutilization of channel capacity. This highlights a key challenge: how to design a codebook that not only supports task-specific inference but also approaches the theoretical limits of channel capacity. To address this challenge, we construct a spectral efficiency-aware codebook design framework that explicitly incorporates the codebook activation probability into the optimization process. Beyond maximizing task performance, we introduce the Wasserstein (WS) distance as a regularization metric to minimize the gap between the learned activation distribution and the optimal channel input distribution. Furthermore, we reinterpret WS theory from a generative perspective to align with the semantic nature of ToSC. Combining the above two aspects, we propose a WS-based adaptive hybrid distribution scheme, termed WS-DC, which learns compact, task-driven and channel-aware latent representations. Experimental results demonstrate that WS-DC not only outperforms existing approaches in inference accuracy but also significantly improves codebook efficiency, offering a promising direction toward capacity-approaching semantic communication systems.

Generative AI (GenAI), exemplified by Large Language Models (LLMs), such as OpenAI's ChatGPT, is revolutionizing various fields. Central to this transformation is Data Center Networking (DCN), which not only provides the infrastructure support for GenAI operation, but also provisions GenAI services to users. Hence, this article explores the interplay between GenAI and DCNs, analyzing their symbiotic relationship and mutual advances. We begin by reviewing the current challenges of DCNs and GenAI-based solutions, such as data augmentation, process automation, and domain transfer. We then discuss the distinctive characteristics of GenAI workloads on DCNs, gaining insights that catalyze the evolution of DCNs to more effectively support GenAI. Moreover, to illustrate the seamless integration of GenAI with DCNs, we present a case study on GenAI-empowered DCN digital twins. Specifically, we employ an LLM equipped with retrieval augmented generation to formulate optimization problems for DCNs (e.g., resource allocation and routing) and adopt diffusion-deep reinforcement learning to solve opti-mization. The experimental results on a representative DCN optimization problem, i.e., knowledge placement, demonstrate the validity and efficiency of our proposals. We anticipate that this article can promote further research to enhance the virtuous interaction between GenAI and DCNs.

This paper proposes a downlink transmission optimization framework for cell-free massive MIMO (CF-mMIMO) systems enhanced by movable antennas (MAs). The proposed architecture significantly expands the spatial degrees of freedom. Moreover, it enhances user-centric cooperation among distributed access points (APs). We formulate a joint optimization problem to maximize the system weighted sum-rate (WSR) by simultaneously optimizing beamforming, user association, and antenna positions. To tackle the problem's non-convexity and coupling, we develop an efficient Joint Beamforming, User Association, and Movable Antenna Position (JBUM) algorithm within a two-loop iterative framework. In the outer loop, dynamic cosine-weighted particle swarm optimization (DCWPSO) is employed to search for optimal antenna locations. In the inner loop, we solve the beamforming subproblem using fractional programming (FP) and address the relaxed user association subproblem using successive convex approximation (SCA). Simulation results demonstrate that, compared with traditional PSO-based optimization, fixed-antenna CF architectures, and single-AP systems, the proposed MA-enhanced scheme achieves performance gains of 5.9%, 21.6%, and 58.8%, respectively. These results validate the effectiveness of integrating movable antennas into CF-mMIMO systems, and highlight its significant advantages in improving both system throughput and optimization flexibility.

The rapid development of emerging applications, including artificial intelligence-generated content (AIGC), has led to a significant demand for computing resources. The crowdsourced edge-cloud system supports this increasing demand by integrating idle resources into elastic resource pools. Within this flexible system, resource providers lease their idle resources for remuneration, while users purchase these resources to support AIGC services. Nevertheless, ensuring fairness, continuity, and efficiency in resource management remains a critical challenge, giving rise to issues such as unfair resource pricing, neglect of resource continuity, and complex strategy optimization. In this paper, we propose Refluid, a resource management framework for dynamic resource flows. To address the challenge of unfair resource pricing, we propose a game-theoretical pricing model to establish a fair and stable guiding transaction price that aligns user demand with provider supply. To address the neglect of continuous resource dynamics, we use a fluid model to characterize resource flows and support dynamic resource management. To address the challenge of complex strategy optimization, we develop a diffusion-based optimization algorithm with associative memory to generate effective management decisions while improving learning efficiency. The evaluation shows that Refluid establishes a fair and stable guiding transaction price that reflects both user demand and provider supply and achieves 3.9%, 10.3%, and 30.8% improvements in social welfare compared to AGOD, ESAC, and IRWS, respectively.

With the advancement of sixth generation (6G) communication technology and artificial intelligent (AI), Electric Intelligent Vehicles (EIVs) have led to a rapid increase in time-bounded and ultra-reliable tasks. However, the growing number of EIVs imposes a high computational complexity on traditional learning-based offloading schemes. This paper proposes a mean field deep reinforcement learning (MFDRL) framework to optimize task offloading decisions aiming to minimize energy consumption for each EIV. Specifically, we first formulate the energy minimization problem as a Markov game while considering the mobility of EIVs, the limited computation resources of EIVs, and the task delay constraints. To reduce the computational complexity of traditional DRL algorithms in deriving the Nash equilibrium in scenarios with a large number of players, we develop a novel DRL algorithm that incorporates mean field theory (MFT), which approximates the joint action space by mean field action. Unlike existing learning-based task offloading approaches, the proposed framework enables each EIV to respond only to the average behavior of the other EIVs, rather than requiring complete knowledge of their individual offloading policies. In this way, the challenge of dimensionality explosion can be effectively mitigated. Simulation results reveal that: 1) the proposed algorithm converges faster than benchmark algorithms, and can achieve superior performance to two benchmark schemes in terms of delay and energy consumption and 2) the proposed algorithm can achieve better scalability than the multi-agent deep deterministic policy gradient (MADDPG) algorithm as the number of TEIVs increases.

This paper introduces DiffCarl, a diffusion-modeled carbon- and risk-aware reinforcement learning algorithm for intelligent operation of multi-microgrid systems. With the growing integration of renewables and increasing system complexity, microgrid communities face significant challenges in real-time energy scheduling and optimization under uncertainty. DiffCarl integrates a diffusion model into a deep reinforcement learning (DRL) framework to enable adaptive energy scheduling under uncertainty and explicitly account for carbon emissions and operational risk. By learning action distributions through a denoising generation process, DiffCarl enhances DRL policy expressiveness and enables carbon- and risk-aware scheduling in dynamic and uncertain microgrid environments. Extensive experimental studies demonstrate that it outperforms classic algorithms and state-of-the-art DRL solutions, with 2.3−30.1

Heterogeneous Networks (HetNets) pose critical challenges for intelligent management due to the diverse user requirements and time-varying wireless conditions. These factors introduce significant decision complexity, which limits the adaptability of existing Deep Reinforcement Learning (DRL) methods. In many DRL algorithms, especially those involving value-based or actor-critic structures, the critic component plays a key role in guiding policy learning by estimating value functions. However, conventional critic models often use shallow architectures that map observations directly to scalar estimates, limiting their ability to handle multi-task complexity. In contrast, recent progress in inference-time scaling of Large Language Models (LLMs) has shown that generating intermediate reasoning steps can significantly improve decision quality. Motivated by this, we propose ReaCritic, a reasoning transformer-based critic-model scaling scheme that brings reasoning-like ability into DRL. ReaCritic performs horizontal reasoning over parallel state-action inputs and vertical reasoning through deep transformer stacks. It is compatible with a broad range of value-based and actor-critic DRL algorithms and enhances generalization in dynamic wireless environments. Extensive experiments demonstrate that ReaCritic improves convergence speed and final performance across various HetNet settings and standard OpenAI Gym control tasks. The code of ReaCritic is available at https://github.com/NICE-HKU/ReaCritic.

Service-based architecture (SBA) has long been recognized as a key enabler of the 5G core network in supporting a wide variety of use cases with challenging requirements, both in the Internet of Things (IoT) and across vertical industries. However, the evolution toward 6G reveals inherent limitations of existing SBA-based core networks, which are primarily designed for static service provisioning. As a consequence, artificial intelligence (AI) is often introduced as an external add-on rather than being natively embedded into the core network architecture, fundamentally limiting the ability to support intelligent, adaptive, and service-aware networking. Furthermore, AI is increasingly becoming a native capability, rather than an external “plug-in” function, in nextgeneration networks. In this context, SBA and AI are natively integrated into an open-source core network (OpenCN). First, we present a novel holistic-service-based OpenCN architecture that incorporates native AI as well as a management and orchestration (MANO) plane. The servicebased MANO plane is designed according to two core principles: (i) decoupling monolithic OpenCN components into independent functions and resources, and (ii) reconfiguring the required functions and resources to construct a customized OpenCN instance tailored to users’ specific requirements. In this manner, each user can be provisioned with a personalized service. Second, we evaluate the performance of the proposed OpenCN framework using a smallscale test network. The results demonstrate that OpenCN achieves higher bandwidth utilization, more efficient resource allocation, and lower energy consumption compared with benchmark solutions. Finally, we discuss several promising research directions for next-generation OpenCNs.

Low-altitude economy (LAE) represents an emerging economic paradigm that redefines commercial and social aerial activities. Large artificial intelligence models (LAIMs) offer transformative potential to further enhance the intelligence of LAE services. However, deploying LAIMs in LAE poses several challenges, including the significant gap between their computational/storage demands and the limited onboard resources of LAE entities, the mismatch between lab-trained LAIMs and dynamic physical environments, and the inefficiencies of traditional decoupled designs for sensing, communication, and computation. To address these issues, we first propose a hierarchical system architecture tailored for LAIM deployment and present representative LAE application scenarios. Next, we explore key enabling techniques that facilitate the mutual co-evolution of LAIMs and low-altitude systems, and introduce a task-oriented execution pipeline for scalable and adaptive service delivery. Then, the proposed framework is validated through real-world case studies. Finally, we outline open challenges to inspire future research.