arXiv subjects

Maryam Hafeez

Publications and source records attributed to Maryam Hafeez.

17 recordsLinked to original sources

arXiv2026

OTel: Building Domain-Specialized Telecom LLM Foundations for Intelligent Networks

Frontier AI models have advanced rapidly, but they still struggle with telecom-specific tasks. We present Open Telco (OTel), an open telecom AI resource with derived datasets for retrieval, reranking, instruction tuning, and safety/abstention, plus 30 full-parameter post-trained baselines across embedding, reranking, and language models. The community has already engaged substantially with the resource: as of May 3, 2026, the released models have been downloaded over 16 million times, and the project has received 157+ pieces of media coverage worldwide. Building on prior open telecom datasets and benchmarks, OTel provides documented telecom data sources, held-out evaluation partitions, trained embedding models, rerankers, context-grounded LLMs, and safety/abstention data in one unified resource. OTel post-training improves performance across all three model families: embedding retrieval reaches 93.5% NDCG@10, reranking reaches 0.952 MRR@10, and language-model correctness reaches 88.2%. We release OTel as a reproducible starting point and invite the community to expand the data, improve embedding and reranking models, and build stronger context-grounded telecom LLMs.

cs.AI↗

arXiv2026

Physically Constrained Federated Additive Models for O-RAN SLA-Risk Prediction

Proactive service assurance in O-RAN requires predicting per-slice SLA violations before they occur. The prediction model must be auditable by operators and must train across base stations without pooling per-slice KPIs, which are commercially sensitive because slices are leased to individual tenants. Neural additive models (NAMs) offer auditability because each KPI contributes through a visible shape function. However, visibility alone does not guarantee physical validity. On the ColO-RAN testbed dataset, unconstrained NAMs learn effects that contradict wireless physics, for example predicting higher risk when channel quality improves. This failure appears under both local and centralized training, and non-IID federated averaging worsens it. We present Monotone FedNAM, a federated additive model in which KPIs with unambiguous physical direction are represented as monotone splines whose constraints survive FedAvg aggregation by construction, while contestable KPIs remain unconstrained. The model trains and operates as a Non-RT RIC rApp and is compact enough for deployment as a Near-RT RIC xApp. Monotone FedNAM eliminates all monotonicity violations, raises constrained shape consistency from 0.71 to 1.00, generalizes to an unseen scheduling policy, and reduces uplink traffic by 65%, at a cost of 0.04 to 0.07 AUC. These results show that physically constrained federated additive models can support auditable SLA risk inference for multi-tenant O-RAN service assurance

cs.LG↗

arXiv2026

Decision-Theoretic Safety Assessment of Persona-Driven Multi-Agent Systems in O-RAN

Autonomous network management in Open Radio Access Networks requires intelligent decision making across conflicting objectives, yet existing LLM based multi agent systems employ homogeneous strategies and lack systematic predeployment validation. We introduce a persona driven multi agent framework where configurable behavioral personas structured specifications encoding optimization priorities, risk tolerance, and decision making style influence five specialized agents (planning, coordination, resource allocation, code generation, analysis). To enable rigorous validation, we develop a three dimensional evaluation framework grounded in decision theory, measuring normative compliance (optimality adherence), prescriptive alignment (behavioral guideline consistency), and behavioral dynamics (emergent system properties). We evaluate 486 persona configurations across two ORAN optimization challenges (energy efficient resource allocation and network load balancing). Results demonstrate that persona agent alignment significantly impacts both individual performance (14.3 percent) and emergent multi agent coordination, with retrieval architecture (GraphRAG vs. RAG) fundamentally constraining customization effectiveness. Single agent persona modifications propagate system wide through cascading effects, with certain combinations exhibiting detectable fundamental incompatibilities. Our framework provides systematic validation mechanisms for deploying LLM based automation in mission critical telecommunications infrastructure.

cs.NI↗

arXiv2026

Vision-Language Models on the Edge for Real-Time Robotic Perception

Vision-Language Models (VLMs) enable multimodal reasoning for robotic perception and interaction, but their deployment in real-world systems remains constrained by latency, limited onboard resources, and privacy risks of cloud offloading. Edge intelligence within 6G, particularly Open RAN and Multi-access Edge Computing (MEC), offers a pathway to address these challenges by bringing computation closer to the data source. This work investigates the deployment of VLMs on ORAN/MEC infrastructure using the Unitree G1 humanoid robot as an embodied testbed. We design a WebRTC-based pipeline that streams multimodal data to an edge node and evaluate LLaMA-3.2-11B-Vision-Instruct deployed at the edge versus in the cloud under real-time conditions. Our results show that edge deployment preserves near-cloud accuracy while reducing end-to-end latency by 5\%. We further evaluate Qwen2-VL-2B-Instruct, a compact model optimized for resource-constrained environments, which achieves sub-second responsiveness, cutting latency by more than half but at the cost of accuracy.

cs.RO↗

arXiv2025

Safety and Risk Pathways in Cooperative Generative Multi-Agent Systems: A Telecom Perspective

Generative multiagent systems are rapidly emerging as transformative tools for scalable automation and adaptive decisionmaking in telecommunications. Despite their promise, these systems introduce novel risks that remain underexplored, particularly when agents operate asynchronously across layered architectures. This paper investigates key safety pathways in telecomfocused Generative MultiAgent Systems (GMAS), emphasizing risks of miscoordination and semantic drift shaped by persona diversity. We propose a modular safety evaluation framework that integrates agentlevel checks on code quality and compliance with systemlevel safety metrics. Using controlled simulations across 32 persona sets, five questions, and multiple iterative runs, we demonstrate progressive improvements in analyzer penalties and AllocatorCoder consistency, alongside persistent vulnerabilities such as policy drift and variability under specific persona combinations. Our findings provide the first domaingrounded evidence that persona design, coding style, and planning orientation directly influence the stability and safety of telecom GMAS, highlighting both promising mitigation strategies and open risks for future deployment.

eess.SY↗

arXiv2025

Agentic AI meets Neural Architecture Search: Proactive Traffic Prediction for AI-RAN

Next-generation wireless networks require intelligent traffic prediction to enable autonomous resource management and handle diverse, dynamic service demands. The Open Radio Access Network (O-RAN) framework provides a promising foundation for embedding machine learning intelligence through its disaggregated architecture and programmable interfaces. This work applies a Neural Architecture Search (NAS)-based framework that dynamically selects and orchestrates efficient Long Short-Term Memory (LSTM) architectures for traffic prediction in O-RAN environments. Our approach leverages the O-RAN paradigm by separating architecture optimisation (via non-RT RIC rApps) from real-time inference (via near-RT RIC xApps), enabling adaptive model deployment based on traffic conditions and resource constraints. Experimental evaluation across six LSTM architectures demonstrates that lightweight models achieve $R^2 \approx 0.91$--$0.93$ with high efficiency for regular traffic, while complex models reach near-perfect accuracy ($R^2 = 0.989$--$0.996$) during critical scenarios. Our NAS-based orchestration achieves a 70-75\% reduction in computational complexity compared to static high-performance models, while maintaining high prediction accuracy when required, thereby enabling scalable deployment in real-world edge environments.

eess.SP↗

arXiv2025

Joint Optimisation of Load Balancing and Energy Efficiency for O-RAN Deployments

Open Radio Access Network (O-RAN) architecture provides an intrinsic capability to exploit key performance monitoring (KPM) within Radio Intelligence Controller (RIC) to derive network optimisation through xApps. These xApps can leverage KPM knowledge to dynamically switch on/off the associated RUs where such a function is supported over the E2 interface. Several existing studies employ artificial intelligence (AI)/Machine Learning (ML) based approaches to realise such dynamic sleeping for increased energy efficiency (EE). Nevertheless, most of these approaches rely upon offloading user equipment (UE) to carve out a sleeping opportunity. Such an approach inherently creates load imbalance across the network. Such load imbalance may impact the throughput performance of offloaded UEs as they might be allocated a lower number of physical resource blocks (PRBs). Maintaining the same PRB allocation while addressing the EE at the network level is a challenging task. To that end, in this article, we present a comprehensive ML-based framework for joint optimisation of load balancing and EE for ORAN deployments. We formulate the problem as a multi-class classification system that predictively evaluates potential RU configurations before optimising the EE, mapping network conditions to three load balance categories (Well Balanced, Moderately Balanced, Imbalanced). Our multi-threshold approach (Conservative, Moderate, Aggressive) accommodates different operational priorities between energy savings and performance assurance. Experimental evaluation using 4.26 million real network measurements from simulations demonstrates that our Random Forest model achieves 98.3% F1-macro performance, representing 195% improvement over traditional baseline strategies.

cs.NI↗

arXiv2025

Semantic-Aware Edge Intelligence for UAV Handover in 6G Networks

6G wireless networks aim to exploit semantic awareness to optimize radio resources. By optimizing the transmission through the lens of the desired goal, the energy consumption of transmissions can also be reduced, and the latency can be improved. To that end, this paper investigates a paradigm in which the capabilities of generative AI (GenAI) on the edge are harnessed for network optimization. In particular, we investigate an Unmanned Aerial Vehicle (UAV) handover framework that takes advantage of GenAI and semantic communication to maintain reliable connectivity. To that end, we propose a framework in which a lightweight MobileBERT language model, fine-tuned using Low-Rank Adaptation (LoRA), is deployed on the UAV. This model processes multi-attribute flight and radio measurements and performs multi-label classification to determine appropriate handover action. Concurrently, the model identifies an appropriate set of contextual "Reason Tags" that elucidate the decision's rationale. Our model, evaluated on a rule-based synthetic dataset of UAV handover scenarios, demonstrates the model's high efficacy in learning these rules, achieving high accuracy in predicting the primary handover decision. The model also shows strong performance in identifying supporting reasons, with an F1 micro-score of approximately 0.9 for reason tags.

eess.SY↗

arXiv2025

Edge Agentic AI Framework for Autonomous Network Optimisation in O-RAN

The deployment of AI agents within legacy Radio Access Network (RAN) infrastructure poses significant safety and reliability challenges for future 6G networks. This paper presents a novel Edge AI framework for autonomous network optimisation in Open RAN environments, addressing these challenges through three core innovations: (1) a persona-based multi-tools architecture enabling distributed, context-aware decision-making; (2) proactive anomaly detection agent powered by traffic predictive tool; and (3) a safety, aligned reward mechanism that balances performance with operational stability. Integrated into the RAN Intelligent Controller (RIC), our framework leverages multimodal data fusion, including network KPIs, a traffic prediction model, and external information sources, to anticipate and respond to dynamic network conditions. Extensive evaluation using realistic 5G scenarios demonstrates that the edge framework achieves zero network outages under high-stress conditions, compared to 8.4% for traditional fixed-power networks and 3.3% for large language model (LLM) agent-based approaches, while maintaining near real-time responsiveness and consistent QoS. These results establish that, when equipped with the right tools and contextual awareness, AI agents can be safely and effectively deployed in critical network infrastructure, laying the framework for intelligent and autonomous 5G and beyond network operations.

eess.SP↗

arXiv2025

Benchmarking Vector, Graph and Hybrid Retrieval Augmented Generation (RAG) Pipelines for Open Radio Access Networks (ORAN)

Generative AI (GenAI) is expected to play a pivotal role in enabling autonomous optimization in future wireless networks. Within the ORAN architecture, Large Language Models (LLMs) can be specialized to generate xApps and rApps by leveraging specifications and API definitions from the RAN Intelligent Controller (RIC) platform. However, fine-tuning base LLMs for telecom-specific tasks remains expensive and resource-intensive. Retrieval-Augmented Generation (RAG) offers a practical alternative through in-context learning, enabling domain adaptation without full retraining. While traditional RAG systems rely on vector-based retrieval, emerging variants such as GraphRAG and Hybrid GraphRAG incorporate knowledge graphs or dual retrieval strategies to support multi-hop reasoning and improve factual grounding. Despite their promise, these methods lack systematic, metric-driven evaluations, particularly in high-stakes domains such as ORAN. In this study, we conduct a comparative evaluation of Vector RAG, GraphRAG, and Hybrid GraphRAG using ORAN specifications. We assess performance across varying question complexities using established generation metrics: faithfulness, answer relevance, context relevance, and factual correctness. Results show that both GraphRAG and Hybrid GraphRAG outperform traditional RAG. Hybrid GraphRAG improves factual correctness by 8%, while GraphRAG improves context relevance by 11%.

cs.AI↗

arXiv2025

EcoFL: Resource Allocation for Energy-Efficient Federated Learning in Multi-RAT ORAN Networks

Federated Learning (FL) enables distributed model training on edge devices while preserving data privacy. However, FL deployments in wireless networks face significant challenges, including communication overhead, unreliable connectivity, and high energy consumption, particularly in dynamic environments. This paper proposes EcoFL, an integrated FL framework that leverages the Open Radio Access Network (ORAN) architecture with multiple Radio Access Technologies (RATs) to enhance communication efficiency and ensure robust FL operations. EcoFL implements a two-stage optimisation approach: an RL-based rApp for dynamic RAT selection that balances energy efficiency with network performance, and a CNN-based xApp for near real-time resource allocation with adaptive policies. This coordinated approach significantly enhances communication resilience under fluctuating network conditions. Experimental results demonstrate competitive FL model performance with 19\% lower power consumption compared to baseline approaches, highlighting substantial potential for scalable, energy-efficient collaborative learning applications.

eess.SP↗

arXiv2025

Proactive AI-and-RAN Workload Orchestration in O-RAN Architectures for 6G Networks

The vision of AI-RAN convergence, as advocated by the AI-RAN Alliance, aims to unlock a unified 6G platform capable of seamlessly supporting AI and RAN workloads over shared infrastructure. However, the architectural framework and intelligent resource orchestration strategies necessary to realize this vision remain largely unexplored. In this paper, we propose a Converged AI-and-ORAN Architectural (CAORA) framework based on O-RAN specifications, enabling the dynamic coexistence of real-time RAN and computationally intensive AI workloads. We design custom xApps within the Near-Real-Time RAN Intelligent Controller (NRT-RIC) to monitor RAN KPIs and expose radio analytics to an End-to-End (E2E) orchestrator via the recently introduced Y1 interface. The orchestrator incorporates workload forecasting and anomaly detection modules, augmenting a Soft Actor-Critic (SAC) reinforcement learning agent that proactively manages resource allocation, including Multi-Instance GPU (MIG) partitioning. Using real-world 5G traffic traces from Barcelona, our trace-driven simulations demonstrate that CAORA achieves near 99\% fulfillment of RAN demands, supports dynamic AI workloads, and maximizes infrastructure utilization even under highly dynamic conditions. Our results reveal that predictive orchestration significantly improves system adaptability, resource efficiency, and service continuity, offering a viable blueprint for future AI-and-RAN converged 6G systems.

cs.NI↗

arXiv2025

From Connectivity to Autonomy: The Dawn of Self-Evolving Communication Systems

This paper envisions 6G as a self-evolving telecom ecosystem, where AI-driven intelligence enables dynamic adaptation beyond static connectivity. We explore the key enablers of autonomous communication systems, spanning reconfigurable infrastructure, adaptive middleware, and intelligent network functions, alongside multi-agent collaboration for distributed decision-making. We explore how these methodologies align with emerging industrial IoT frameworks, ensuring seamless integration within digital manufacturing processes. Our findings emphasize the potential for improved real-time decision-making, optimizing efficiency, and reducing latency in networked control systems. The discussion addresses ethical challenges, research directions, and standardization efforts, concluding with a technology stack roadmap to guide future developments. By leveraging state-of-the-art 6G network management techniques, this research contributes to the next generation of intelligent automation solutions, bridging the gap between theoretical advancements and real-world industrial applications.

eess.SY↗

arXiv2025

FedORA: Resource Allocation for Federated Learning in ORAN using Radio Intelligent Controllers

This work proposes an integrated approach for optimising Federated Learning (FL) communication in dynamic and heterogeneous network environments. Leveraging the modular flexibility of the Open Radio Access Network (ORAN) architecture and multiple Radio Access Technologies (RATs), we aim to enhance data transmission efficiency and mitigate client-server communication constraints within the FL framework. Our system employs a two-stage optimisation strategy using ORAN's rApps and xApps. In the first stage, Reinforcement Learning (RL) based rApp is used to dynamically select each user's optimal Radio Access Technology (RAT), balancing energy efficiency with network performance. In the second stage, a model-based xApp facilitates near-real-time resource allocation optimisation through predefined policies to achieve optimal network performance. The dynamic RAT selection and resource allocation capabilities enabled by ORAN and multi-RAT contribute to robust communication resilience in dynamic network environments. Our approach demonstrates competitive performance with low power consumption compared to other state-of-the-art models, showcasing its potential for real-time applications demanding both accuracy and efficiency. This robust and comprehensive framework, enabling clients to utilise available resources effectively, highlights the potential for scalable, collaborative learning applications prioritising energy efficiency and network performance.

eess.SY↗

arXiv2025

The Interplay of AI-and-RAN: Dynamic Resource Allocation for Converged 6G Platform

The concept of AI-RAN as specified by the AI-RAN alliance is geared to explore a converged 6G platform that can support management, orchestration, and deployment of both AI and RAN workloads. This concept is central to the development of a 6G architecture that aims to exploit the accelerated compute capabilities for supporting both real-time signal processing and offloading of Generative AI (GenAI) workloads. However, both the architectural framework required to support this vision and the dynamic resource allocation strategy are still in their infancy. The O-RAN architecture intrinsically allows cloud-native disaggregated implementation. Consequently, we explore a framework that can allow orchestration of AI-and-RAN workloads by expanding the Near Real-Time RAN Intelligent Controller (NRT-RIC) within O-RAN. The framework incorporates a monitoring xApp that tracks RAN KPIs and exposes radio analytics to the proposed E2E orchestrator via a recently introduced Y1 interface. The orchestrator implements a Soft Actor-Critic (SAC) reinforcement learning algorithm to dynamically allocate critical computing resources, e.g., Multi-Instance GPUs (MIGs), between latency-sensitive RAN network functions and computationally intensive AI workloads on shared RAN infrastructure. The proposed framework provides insight on how the traditional RAN architecture can be evolved to inherently support emerging GenAI workloads. Our framework prioritizes the real-time requirements of RAN workloads while maintaining efficient resource sharing for AI applications. The simulation results demonstrate the benefits of the proposed framework, as it meets nearly 99% of the requests for RAN workload while effectively supporting AI workloads and achieving 100% utilization of the RAN infrastructure resources in a dynamic environment.

cs.NI↗

arXiv2025

Large-Scale AI in Telecom: Charting the Roadmap for Innovation, Scalability, and Enhanced Digital Experiences

This white paper discusses the role of large-scale AI in the telecommunications industry, with a specific focus on the potential of generative AI to revolutionize network functions and user experiences, especially in the context of 6G systems. It highlights the development and deployment of Large Telecom Models (LTMs), which are tailored AI models designed to address the complex challenges faced by modern telecom networks. The paper covers a wide range of topics, from the architecture and deployment strategies of LTMs to their applications in network management, resource allocation, and optimization. It also explores the regulatory, ethical, and standardization considerations for LTMs, offering insights into their future integration into telecom infrastructure. The goal is to provide a comprehensive roadmap for the adoption of LTMs to enhance scalability, performance, and user-centric innovation in telecom networks.

cs.NI↗

arXiv2024

Generative AI on the Edge: Architecture and Performance Evaluation

6G's AI native vision of embedding advance intelligence in the network while bringing it closer to the user requires a systematic evaluation of Generative AI (GenAI) models on edge devices. Rapidly emerging solutions based on Open RAN (ORAN) and Network-in-a-Box strongly advocate the use of low-cost, off-the-shelf components for simpler and efficient deployment, e.g., in provisioning rural connectivity. In this context, conceptual architecture, hardware testbeds and precise performance quantification of Large Language Models (LLMs) on off-the-shelf edge devices remains largely unexplored. This research investigates computationally demanding LLM inference on a single commodity Raspberry Pi serving as an edge testbed for ORAN. We investigate various LLMs, including small, medium and large models, on a Raspberry Pi 5 Cluster using a lightweight Kubernetes distribution (K3s) with modular prompting implementation. We study its feasibility and limitations by analyzing throughput, latency, accuracy and efficiency. Our findings indicate that CPU-only deployment of lightweight models, such as Yi, Phi, and Llama3, can effectively support edge applications, achieving a generation throughput of 5 to 12 tokens per second with less than 50\% CPU and RAM usage. We conclude that GenAI on the edge offers localized inference in remote or bandwidth-constrained environments in 6G networks without reliance on cloud infrastructure.

cs.DC↗