Recent advances in large language models (LLMs) have driven remarkable progress across diverse NLP benchmarks. However, the application of these models to sophisticated domains such as wireless communications raises new challenges. This paper addresses the evaluation gap for LLMs in telecommunications by introducing Last Dance of Telecommunications (LDOT) – a comprehensive benchmark suite for telecom-related tasks. LDOT encompasses a broad range of problem categories, including conceptual telecom questions, mathematical and logical reasoning problems, and complex network optimization scenarios, specifically designed to challenging LLMs’ high-level reasoning and domain-specific knowledge. Using LDOT, we rigorously assess state-of-the-art (SOTA) LLMs (both closed-source and open-source) on domain-specific tasks. Our results reveal that while general-purpose LLMs exhibit strong performance on basic telecom knowledge questions, they struggle with reasoning-intensive wireless problems. Notably, certain multi-step optimization and planning tasks in LDOT remain unsolved by even the best models, exposing performance gaps that are not apparent from existing saturated benchmarks. We provide a detailed failure analysis to pinpoint whether these limitations arise from insufficient telecom-specific knowledge or from inadequate reasoning capabilities. The LDOT dataset and our evaluation findings aim to facilitate the development of more robust domain-adapted LLMs for next-generation wireless communications.

This paper addresses the challenge of training federated learning (FL) algorithms over practical wireless networks enhanced by non-orthogonal multiple access (NOMA). In FL procedures, model parameters are transmitted over wireless links, and factors such as packet errors can significantly impact training quality, especially with non-independent and identically distributed (non-IID) datasets. To address this issue, we formulate the complex learning and wireless resource allocation problem as an optimization task aimed at minimizing an FL loss function, thereby capturing the performance of the FL algorithm. Under mild assumptions, we derive the expected convergence rate of the FL algorithm, quantifying the impact of wireless factors on FL. Leveraging this convergence analysis, we determine the optimal transmit power for each user through quadratic transform (QT) and a proposed low-complexity iterative method with successive convex approximation (SCA) serving as benchmark. Extensive simulation results demonstrate that the proposed scheme significantly outperforms traditional schemes from both the optimization and FL performance perspectives.

Large language models (LLMs) have shown great promise in many domains, yet their potential to transform wireless communications, where the escalating complexity of the network outpaces traditional model-based methods, remains largely untapped. Addressing this gap is critical for the next generation of intelligent and adaptive 6G systems. In this work, we develop a specialized dataset aimed at enhancing the evaluation and fine-tuning of LLMs specifically for wireless communication applications. The dataset includes a diverse set of multi-hop questions, including true/false and multiple-choice types, spanning varying difficulty levels from easy to hard. By utilizing advanced language models for entity extraction and question generation, rigorous data curation processes are employed to maintain high quality and relevance. Additionally, we introduce a Pointwise V-Information (PVI) based fine-tuning method, providing a detailed theoretical analysis and justification for its use in quantifying the information content of training data with 2.24% and 1.31% performance boost for different models compared to baselines, respectively. To demonstrate the effectiveness of the fine-tuned models with the proposed methodologies on practical tasks, we also consider different tasks, including summarizing optimization problems from technical papers and solving the mathematical problems related to non-orthogonal multiple access (NOMA), which are generated by using the proposed multi-agent framework. Simulation results show significant performance gain in summarization tasks with 20.9% in the ROUGE-L metrics. We also study the scaling laws of fine-tuning LLMs and the challenges LLMs face in the field of wireless communications, offering insights into their adaptation to wireless communication tasks. This dataset and fine-tuning methodology aim to enhance the training and evaluation of LLMs, contributing to advancements in LLMs for wireless communication research and applications.

The combination of non-orthogonal multiple access (NOMA) and mobile edge computing (MEC) has recently received great attention for maximizing energy efficiency (EE) in wireless networks. Previous studies mainly focus on single-ratio EE maximization using pure orthogonal multiple access (OMA) or NOMA scheme in MEC networks. This paper proposes a dynamic hybrid NOMA scheme that dynamically transforms between pure OMA and pure NOMA under the general network status in the multi-user MEC network. An offloading multi-ratio and non-convex EE problem is formulated. Two iterative algorithms, i.e., the successive Dinkelbach method-based algorithm and the closed-form nested fractional programming (FP) algorithm, are proposed to efficiently solve the offloading EE over multiple users’ time and power within polynomial time based on different FP techniques, which provides feasible ideas to solve such multi-ratio optimization problems. The Dinkelbach method-based algorithm decomposes and optimizes the offloading EE user-wise. The nested FP algorithm jointly maximizes multiple users’ EE by transforming the multi-ratio problem into equivalent linear forms, where the closed-form power solution is obtained. Simulation results show the superiority of the dynamic hybrid NOMA strategy and the feasibility of the two proposed algorithms. Critical insights are obtained over the dynamic hybrid NOMA strategies in the general MEC network.

Although Federated Learning (FL) has garnered increasing attention from researchers, the development of ad-vanced FL frameworks incorporating multiple access techniques remains relatively underexplored. This paper investigates the integration of a novel clustered federated learning (CFL) framework with non-orthogonal multiple access (NOMA) in environments with non-independent and identically distributed (non-iid) datasets. To explore the potential benefits of the proposed framework, the optimization problem is formulated as an energy minimization problem, which includes sub-channel and power allocation. The formulated problem is divided into two sub-problems, respectively solved by the matching-based algorithm and Karush-Kuhn-Tucker (KKT) conditions, in which the closed-form solution is derived. Our simulation results demon-strate that jointly optimizing sub-channel and power allocation in NOMA-enhanced networks can lead to a significant improvement in test accuracy and convergence speed in the proposed FL framework.

This study explores the benefits of integrating the novel clustered federated learning (CFL) approach with non-orthogonal multiple access (NOMA) under non-independent and identically distributed (non-IID) datasets, where multiple devices participate in the aggregation with time limitations and a finite number of sub-channels. A detailed theoretical analysis of the generalization gap that measures the degree of non-IID in the data distribution is presented. Following that, solutions to address the challenges posed by non-IID conditions are proposed with the analysis of the properties. Specifically, users’ data distributions are parameterized as concentration parameters and grouped using spectral clustering, with Dirichlet distribution serving as the prior. The investigation into the generalization gap and convergence rate guides the design of sub-channel assignments through the matching-based algorithm, and the power allocation is achieved by Karush-Kuhn-Tucker (KKT) conditions with the derived closed-form solution. The extensive simulation results show that the proposed cluster-based FL framework can outperform FL baselines in terms of both test accuracy and convergence rate. Moreover, jointly optimizing sub-channel and power allocation in NOMA-enhanced networks can lead to a significant improvement.