Mohamed Sana

Mohamed Sana, Emilio Calvanese Strinati

We consider a multi-user semantic communications system in which agents (transmitters and receivers) interact through the exchange of semantic messages to convey meanings. In this context, languages are instrumental in structuring the construction and consolidation of knowledge, influencing conceptual representation and semantic extraction and interpretation. Yet, the crucial role of languages in semantic communications is often overlooked. When this is not the case, agent languages are assumed compatible and unambiguously interoperable, ignoring practical limitations that may arise due to language mismatching. This is the focus of this work. When agents use distinct languages, message interpretation is prone to semantic noise resulting from critical distortion introduced by semantic channels. To address this problem, this paper proposes a new semantic channel equalizer to counteract and limit the critical ambiguity in message interpretation. Our proposed solution models the mismatch of languages with measurable transformations over semantic representation spaces. We achieve this using optimal transport theory, where we model such transformations as transportation maps. Then, to recover at the receiver the meaning intended by the teacher we operate semantic equalization to compensate for the transformation introduced by the semantic channel, either before transmission and/or after the reception of semantic messages. We implement the proposed approach as an operation over a codebook of transformations specifically designed for successful communication. Numerical results show that the proposed semantic channel equalizer outperforms traditional approaches in terms of operational complexity and transmission accuracy.

Mohamed Sana, Nicola Piovesan, Antonio De Domenico et al.

We investigate a narrow but common failure mode of GRPO-style reinforcement learning in the context of sparse verifiable rewards: early updates contain more responses with negative advantages than those with positive advantages, while response-level length normalization ties the magnitude of the update to the length of the output. We propose Hysteretic Policy Optimization (HPO), a minimal modification of GRPO that reduces the weight of negative-advantage updates and replaces per-response length normalization with mean-length normalization. We further introduce Adaptive HPO (A-HPO), which sets the hysteretic weight based on batch-level advantage-sign statistics, thereby removing the need for tuning a fixed hysteretic weight. In our TeleLogs and Countdown experiments, A-HPO improves the reward per update compared to GRPO, with the largest gains in early sparse reward regimes. On TeleLogs, A-HPO achieves a final reward of 0.84, outperforming SAPO by 5%, GSPO by 11%, and GRPO by 15%, while maintaining a comparable response-length. On Countdown, A-HPO achieves the largest gains in initial and most difficult configurations across 1.5B-7B models. Ablation studies on the hysteretic weight show that the gains of A-HPO come from better balancing the contributions of positive and negative advantages compared to positive-only or fully symmetric updates.

Esteban Catté, Mohamed Sana, Mickael Maman

This paper addresses the efficient management of Mobile Access Points (MAPs), which are Unmanned Aerial Vehicles (UAV), in 5G networks. We propose a two-level hierarchical architecture, which dynamically reconfigures the network while considering Integrated Access-Backhaul (IAB) constraints. The high-layer decision process determines the number of MAPs through consensus, and we develop a joint optimization process to account for co-dependence in network self-management. In the low-layer, MAPs manage their placement using a double-attention based Deep Reinforcement Learning (DRL) model that encourages cooperation without retraining. To improve generalization and reduce complexity, we propose a federated mechanism for training and sharing one placement model for every MAP in the low-layer. Additionally, we jointly optimize the placement and backhaul connectivity of MAPs using a multi-objective reward function, considering the impact of varying MAP placement on wireless backhaul connectivity.

Vincenzo Colle, Mohamed Sana, Nicola Piovesan et al.

The increasing adoption of artificial intelligence in telecommunications has raised interest in the capability of Large Language Models (LLMs) to address domain-specific, mathematically intensive tasks. Although recent advancements have improved the performance of LLMs in general mathematical reasoning, their effectiveness within specialized domains, such as signal processing, network optimization, and performance analysis, remains largely unexplored. To address this gap, we introduce TeleMath, the first benchmark dataset specifically designed to evaluate LLM performance in solving mathematical problems with numerical solutions in the telecommunications domain. Comprising 500 question-answer (QnA) pairs, TeleMath covers a wide spectrum of topics in the telecommunications field. This paper outlines the proposed QnAs generation pipeline, starting from a selected seed of problems crafted by Subject Matter Experts. The evaluation of a wide range of open-source LLMs reveals that best performance on TeleMath is achieved by recent models explicitly designed for mathematical or logical reasoning. In contrast, general-purpose models, even those with a large number of parameters, often struggle with these challenges. We have released the dataset and the evaluation code to ease result reproducibility and support future research.

Mohamed Sana, Nicola Piovesan, Antonio De Domenico et al.

Root Cause Analysis (RCA) in mobile networks remains a challenging task due to the need for interpretability, domain expertise, and causal reasoning. In this work, we propose a lightweight framework that leverages Large Language Models (LLMs) for RCA. To do so, we introduce TeleLogs, a curated dataset of annotated troubleshooting problems designed to benchmark RCA capabilities. Our evaluation reveals that existing open-source reasoning LLMs struggle with these problems, underscoring the need for domain-specific adaptation. To address this issue, we propose a two-stage training methodology that combines supervised fine-tuning with reinforcement learning to improve the accuracy and reasoning quality of LLMs. The proposed approach fine-tunes a series of RCA models to integrate domain knowledge and generate structured, multi-step diagnostic explanations, improving both interpretability and effectiveness. Extensive experiments across multiple LLM sizes show significant performance gains over state-of-the-art reasoning and non-reasoning models, including strong generalization to randomized test variants. These results demonstrate the promise of domain-adapted, reasoning-enhanced LLMs for practical and explainable RCA in network operation and management.

Tomás Hüttebräucker, Simone Fiorellino, Mohamed Sana et al.

In multi-user semantic communication, language mismatche poses a significant challenge when independently trained agents interact. We present a novel semantic equalization algorithm that enables communication between agents with different languages without additional retraining. Our algorithm is based on relative representations, a framework that enables different agents employing different neural network models to have unified representation. It proceeds by projecting the latent vectors of different models into a common space defined relative to a set of data samples called \textit{anchors}, whose number equals the dimension of the resulting space. A communication between different agents translates to a communication of semantic symbols sampled from this relative space. This approach, in addition to aligning the semantic representations of different agents, allows compressing the amount of information being exchanged, by appropriately selecting the number of anchors. Eventually, we introduce a novel anchor selection strategy, which advantageously determines prototypical anchors, capturing the most relevant information for the downstream task. Our numerical results show the effectiveness of the proposed approach allowing seamless communication between agents with radically different models, including differences in terms of neural network architecture and datasets used for initial training.

Anas Ezzakri, Nicola Piovesan, Mohamed Sana et al.

Language Models (LLMs) are increasingly explored in the telecom industry to support engineering tasks, accelerate troubleshooting, and assist in interpreting complex technical documents. However, recent studies show that LLMs perform poorly on telecom standards, particularly 3GPP specifications. We argue that a key reason is that these standards densely include tables to present essential information, yet the LLM knowledge and interpretation ability of such tables remains largely unexamined. To address this gap, we introduce TeleTables, a benchmark designed to evaluate both the implicit knowledge LLMs have about tables in technical specifications and their explicit ability to interpret them. TeleTables is built through a novel multi-stage data generation pipeline that extracts tables from 3GPP standards and uses multimodal and reasoning-oriented LLMs to generate and validate questions. The resulting dataset, which is publicly available, comprises 500 human-verified question-answer pairs, each associated with the corresponding table in multiple formats. Our evaluation shows that, smaller models (under 10B parameters) struggle both to recall 3GPP knowledge and to interpret tables, indicating the limited exposure to telecom standards in their pretraining and the insufficient inductive biases for navigating complex technical material. Larger models, on the other hand, show stronger reasoning on table interpretation. Overall, TeleTables highlights the need for domain-specialized fine-tuning to reliably interpret and reason over telecom standards.

Dmitry Akulov, Mohamed Sana, Antonio De Domenico et al.

Large language models (LLMs) rely on key-value (KV) caches for efficient autoregressive decoding; however, cache size grows linearly with context length and model depth, becoming a major bottleneck in long-context inference. Prior KV cache compression methods either enforce rigid heuristics, disrupt tensor layouts with per-attention-head variability, or require specialized compute kernels. We propose a simple, yet effective, KV cache compression framework based on attention-guided, layer-adaptive composite tokens. Our method aggregates attention scores to estimate token importance, selects head-specific tokens independently, and aligns them into composite tokens that respect the uniform cache structure required by existing inference engines. A global allocation mechanism further adapts retention budgets across layers, assigning more capacity to layers with informative tokens. This approach achieves significant memory reduction while preserving accuracy, consistently outperforming prior structured and semi-structured methods. Crucially, our approach remains fully compatible with standard inference pipelines, offering a practical and scalable solution for efficient long-context LLM deployment.

Luca-Andrei Fechete, Mohamed Sana, Fadhel Ayed et al.

Conventional time-series forecasting methods typically aim to minimize overall prediction error, without accounting for the varying importance of different forecast ranges in downstream applications. We propose a training methodology that enables forecasting models to adapt their focus to application-specific regions of interest at inference time, without retraining. The approach partitions the prediction space into fine-grained segments during training, which are dynamically reweighted and aggregated to emphasize the target range specified by the application. Unlike prior methods that predefine these ranges, our framework supports flexible, on-demand adjustments. Experiments on standard benchmarks and a newly collected wireless communication dataset demonstrate that our method not only improves forecast accuracy within regions of interest but also yields measurable gains in downstream task performance. These results highlight the potential for closer integration between predictive modeling and decision-making in real-world systems.

Tomás Hüttebräucker, Mohamed Sana, Emilio Calvanese Strinati

We relax the constraint of a shared language between agents in a semantic and goal-oriented communication system to explore the effect of language mismatch in distributed task solving. We propose a mathematical framework, which provides a modelling and a measure of the semantic distortion introduced in the communication when agents use distinct languages. We then propose a new approach to semantic channel equalization with proven effectiveness through numerical evaluations.

Tomás Hüttebräucker, Mohamed Sana, Emilio Calvanese Strinati

In forthcoming AI-assisted 6G networks, integrating semantic, pragmatic, and goal-oriented communication strategies becomes imperative. This integration will enable sensing, transmission, and processing of exclusively pertinent task data, ensuring conveyed information possesses understandable, pragmatic semantic significance, aligning with destination needs and goals. Without doubt, no communication is error free. Within this context, besides errors stemming from typical wireless communication dynamics, potential distortions between transmitter-intended and receiver-interpreted meanings can emerge due to limitations in semantic processing capabilities, as well as language and knowledge representation disparities between transmitters and receivers. The main contribution of this paper is two-fold. First, it proposes and details a novel mathematical modeling of errors stemming from language mismatches at both semantic and effectiveness levels. Second, it provides a novel algorithmic solution to counteract these types of errors which leverages optimal transport theory. Our numerical results show the potential of the proposed mechanism to compensate for language mismatches, thereby enhancing the attainability of reliable communication under noisy communication environments.

Mohamed Sana, Emilio Calvanese Strinati

Recently, semantic communications are envisioned as a key enabler of future 6G networks. Back to Shannon's information theory, the goal of communication has long been to guarantee the correct reception of transmitted messages irrespective of their meaning. However, in general, whenever communication occurs to convey a meaning, what matters is the receiver's understanding of the transmitted message and not necessarily its correct reconstruction. Hence, semantic communications introduce a new paradigm: transmitting only relevant information sufficient for the receiver to capture the meaning intended can save significant communication bandwidth. Thus, this work explores the opportunity offered by semantic communications for beyond 5G networks. In particular, we focus on the benefit of semantic compression. We refer to semantic message as a sequence of well-formed symbols learned from the "meaning" underlying data, which have to be interpreted at the receiver. This requires a reasoning unit, here artificial, on a knowledge base: a symbolic knowledge representation of the specific application. Therefore, we present and detail a novel architecture that enables representation learning of semantic symbols for effective semantic communications. We first discuss theoretical aspects and successfully design objective functions, which help learn effective semantic encoders and decoders. Eventually, we show promising numerical results for the scenario of text transmission, especially when the sender and receiver speak different languages.

Mohamed Sana, Nicola di Pietro, Emilio Calvanese Strinati

We study the problem of user association, namely finding the optimal assignment of user equipment to base stations to achieve a targeted network performance. In this paper, we focus on the knowledge transferability of association policies. Indeed, traditional non-trivial user association schemes are often scenario-specific or deployment-specific and require a policy re-design or re-learning when the number or the position of the users change. In contrast, transferability allows to apply a single user association policy, devised for a specific scenario, to other distinct user deployments, without needing a substantial re-learning or re-design phase and considerably reducing its computational and management complexity. To achieve transferability, we first cast user association as a multi-agent reinforcement learning problem. Then, based on a neural attention mechanism that we specifically conceived for this context, we propose a novel distributed policy network architecture, which is transferable among users with zero-shot generalization capability i.e., without requiring additional training.Numerical results show the effectiveness of our solution in terms of overall network communication rate, outperforming centralized benchmarks even when the number of users doubles with respect to the initial training point.

Mohamed Sana, Mattia Merluzzi, Nicola di Pietro et al.

In this work, we study the problem of energy-efficient computation offloading enabled by edge computing. In the considered scenario, multiple users simultaneously compete for limited radio and edge computing resources to get offloaded tasks processed under a delay constraint, with the possibility of exploiting low power sleep modes at all network nodes. The radio resource allocation takes into account inter- and intra-cell interference, and the duty cycles of the radio and computing equipment have to be jointly optimized to minimize the overall energy consumption. To address this issue, we formulate the underlying problem as a dynamic long-term optimization. Then, based on Lyapunov stochastic optimization tools, we decouple the formulated problem into a CPU scheduling problem and a radio resource allocation problem to be solved in a per-slot basis. Whereas the first one can be optimally and efficiently solved using a fast iterative algorithm, the second one is solved using distributed multi-agent reinforcement learning due to its non-convexity and NP-hardness. The resulting framework achieves up to 96.5% performance of the optimal strategy based on exhaustive search, while drastically reducing complexity. The proposed solution also allows to increase the network's energy efficiency compared to a benchmark heuristic approach.

Mohamed Sana, Antonio De Domenico, Wei Yu et al.

Network densification and millimeter-wave technologies are key enablers to fulfill the capacity and data rate requirements of the fifth generation (5G) of mobile networks. In this context, designing low-complexity policies with local observations, yet able to adapt the user association with respect to the global network state and to the network dynamics is a challenge. In fact, the frameworks proposed in literature require continuous access to global network information and to recompute the association when the radio environment changes. With the complexity associated to such an approach, these solutions are not well suited to dense 5G networks. In this paper, we address this issue by designing a scalable and flexible algorithm for user association based on multi-agent reinforcement learning. In this approach, users act as independent agents that, based on their local observations only, learn to autonomously coordinate their actions in order to optimize the network sum-rate. Since there is no direct information exchange among the agents, we also limit the signaling overhead. Simulation results show that the proposed algorithm is able to adapt to (fast) changes of radio environment, thus providing large sum-rate gain in comparison to state-of-the-art solutions.