arXiv subjects
Publications and source records attributed to Marco Ruffini.
Challenges associated with densification of radio access networks are motivating exploration of more efficient and scalable architectures. We examine recent progress in one direction that involves moving beyond radio and optical convergence to full confluence.
We investigate how quantum computers can be used to emulate quantum networks and study their performance under practical impairments. In particular, we evaluate how degraded entanglement and communication latency affect teleportation-based distributed multipartite-entanglement-state construction. We model imperfect Bell-pair sources using depolarizing noise channels and classical communication delays using thermal relaxation. We implement the depolarization using Stinespring dilation, randomly applied Pauli errors, and quasi-probability decompositions, evaluating the latter two on IQM quantum hardware and all three in simulation. We then study the performance of the entanglement distribution under noise generated by the aforementioned models. Although these noise models are mathematically equivalent, we find that hardware constraints result in profound differences in the corresponding results, highlighting the importance of careful experiment design.
We present a DLM-anchored hybrid physics/ML framework for brownfield optical links that accurately predicts per-channel power, OSNR, and GSNR. Calibrating span/ILA boundaries via DLM yields OSNR/GSNR errors of no more than 0.39/0.43 dB across single-channel and OSaaS provisioning.
We present a fully unsupervised Fast-Slow DSVDD detector for continuous State-of-Polarization monitoring on a deployed subsea cable. Trained without event labels, it ranks all five confirmed trawler contacts within the top 13 of 122,174 recordings and surfaces additional corroborated cable-contact events.
We present an ML-based vessel detection and localization system, trained with weak supervision from imperfect AIS labels, that achieves a 97.8% detection rate at 1.98% false-trigger rate, successfully identifies dark-vessel events from unlabeled data.
We present a case study on the Emerald Fibre Bridge Link, an operational subsea telecom cable connecting Dublin and North Wales, examining DAS vessel-related signatures jointly with concurrent AIS data. The observations show that vessel-related DAS responses depend on local cable sensitivity and background conditions, while their interpretation is complicated by imperfect AIS reporting. Examining vessel-crossing events jointly, we identify representative DAS-AIS association patterns, ranging from clear vessel matches to offset, ambiguous, AIS-incomplete, AIS-silent-candidate, and non-vessel confounders. These observations reveal the gap between physical measurements at the cable and cooperative vessel reporting, providing practical insights for designing future DAS-assisted cable-protection workflows.
Continual adaptation is essential for multimodal large language models (MLLMs) deployed across evolving domains, but the state-of-the-art MR-LoRA method highly relies on the assumption that a MLLM-based router is necessary to process complex multimodal inputs. This paper revisits this claim on the MLLM-CL benchmark and argues for two claims. \textbf{First}, routing does not require an MLLM: a simple training-free, replay-free ptotypical routing method (\textsc{RePRo}), uses frozen pretrained features and task prototypes to match the MLLM-based router of MR-LoRA at far lower computational cost. \textbf{Second}, shared experts do not improve continual learning for MLLMs, despite their theoretical appeal. We show that these findings arise from two structural limitations of MLLM-CL: (1) its tasks are \textbf{highly separable} in representation space, and (2) its fixed task order makes conclusions \textbf{sensitive to a single curriculum} rather than robust across diverse continual-learning trajectories. As a result, the benchmark primarily rewards learning in isolation rather than genuine continual transfer. This motivates a new design for future benchmarks of continual MLLM learning, with overlapping task manifolds, multiple task orders, fine-grained domain shifts, and evaluation protocols that reward forward transfer as well as retention.
We develop an adaptive OFDM framework for underwater acoustic communications based on PatchCSI-T, a Transformer-based multistep channel prediction model with feature-independent modeling and parameter sharing. Combined with a greedy adaptive modulation and power allocation scheme, the proposed approach enables accurate, low-latency CSI forecasting and improves end-to-end BER and spectral efficiency on real-world UWA channel datasets.
Heterogeneous LoRA-rank methods address system heterogeneity in federated fine-tuning of foundation models by assigning client-specific ranks based on computational capabilities. However, these methods achieve only marginal computational savings, as dense feed-forward computations dominate. Sparse Mixture-of-Experts (SMoE) provides a promising alternative through conditional computation, yet we identify that its naive application to heterogeneous federated settings introduces two critical discordances: (i) expert utilization imbalance and (ii) non-differentiability of Top-K routing. Our convergence analysis demonstrates that these discordances lead to degraded convergence, particularly for resource-constrained clients. To address these challenges, we propose Universally Balanced Sparse Mixture-of-Experts (UB-SMoE), which introduces Dynamic Modulated Routing (DMR) to rebalance expert utilization, and Universal Pseudo-Gradient (PG) to reconstruct learning signals for non-activated experts. These mechanisms form a self-reinforcing cycle that maintains expert viability across heterogeneous clients. Experiments on benchmarks show that UB-SMoE achieves up to $45.0\%$ computational reduction on low-resource clients while improving their performance by $8.7 \times$ compared to existing heterogeneous LoRA-rank methods.
Root cause analysis (RCA) in complex systems is challenging due to error propagation across multiple variables, the need for structural causal knowledge, and the computational cost of inference at test time. We introduce PRIM (Prior-fitted Root cause Identification with Meta-learning), a causal meta-learning approach that frames RCA as a Bayesian inference task over a synthetic prior of causal models. By marginalising out structural uncertainty, PRIM implicitly identifies changes in the data-generating mechanism between baseline and anomalous periods. In doing so, PRIM infers distributional differences without explicit statistical testing, and implicitly learns causal structure without model fitting at test time. Following the simulation-based meta-learning paradigm of prior-fitted networks, PRIM uses a Model-Averaged Causal Estimation (MACE) transformer neural process that jointly attends over observational and anomalous samples and the causal structure of nodes, enabling zero-shot inference in 17,ms for systems with up to 100 variables. Across synthetic benchmarks and two realistic benchmark datasets, PetShop and CausRCA, PRIM is competitive with methods that are aware of the system's causal graphical structure a priori while outperforming graph-unaware methods on several tasks. Lightweight fine-tuning to specific domains and data dynamics improves performance further.
Machine learning (ML) models trained to detect physical-layer threats on one optical fiber system often fail catastrophically when applied to a different system, due to variations in operating wavelength, fiber properties, and network architecture. To overcome this, we propose a Domain Adaptation (DA) framework based on a Variational Autoencoder (VAE) that learns a shared representation capturing event signatures common to both systems while suppressing system-specific differences. The shared encoder is first trained on the combined data from two distinct optical systems: a 21 km O-band dark-fiber testbed (System 1) and a 63.4 km C-band live metro ring (System 2). The encoder is then frozen, and a classifier is trained using labels from an individual system. The proposed approach achieves 95.3% and 73.5% cross-system accuracy when moving from System 1 to System 2 and vice versa, respectively. This corresponds to gains of 83.4% and 51% over a fully supervised Deep Neural Network (DNN) baseline trained on a single system, while preserving intra-system performance.
We propose a spectrum-configuration framework for open and disaggregated optical systems that maximizes throughput while guaranteeing the quality of transmission (QoT) margins. The framework jointly optimizes transceiver parameters, including modulation format, symbol rate, pulse-shaping roll-off factor, and wavelength-selective switch (WSS) bandwidth, under fixed spectral allocation constraints. The impact of roll-off factor optimization is first experimentally evaluated in the presence of cascaded WSS filtering, demonstrating measurable QoT gains for both single- and multi-channel transmission. Building on these observations, a knapsack-based optimization is applied in the context of Optical Spectrum as a Service (OSaaS) to select service configurations that maximize aggregate throughput within a fixed spectrum width and limited transceiver resources. Experimental validation on a metro-scale open testbed confirms the effectiveness of the proposed approach in achieving efficient spectrum utilization and adaptive throughput-margin trade-offs.
Time series foundation models (TSFMs) require diverse, real-world datasets to adapt across varying domains and temporal frequencies. However, current large-scale datasets predominantly focus on low-frequency time series with sampling intervals, i.e., time resolution, in the range of seconds to years, hindering their ability to capture the nuances of high-frequency time series data. To address this limitation, we introduce a novel dataset, \textbf{msData}, that captures millisecond-resolution wireless and traffic conditions from an operational 5G wireless deployment, expanding the scope of TSFMs to incorporate high-frequency data for pre-training. Further, the dataset introduces a new domain, namely, wireless networks, thus complementing existing more general domains like energy and finance. The dataset also provides use cases for short-term forecasting, with prediction horizons spanning from 1 millisecond (1 step) to 96 milliseconds (96 steps). To demonstrate the utility of the dataset, we benchmark traditional machine learning models, transformer-based deep learning models, and TSFMs on forecasting tasks using representative subsets of the data, including a static mobility pattern within YouTube traffic class and a train mobility pattern within Web Browsing traffic class. Across these data distributions, we demonstrate that most TSFM model configurations perform poorly in both zero-shot and fine-tuned settings. Our work underscores the importance of incorporating high-frequency datasets during pre-training and forecasting to enhance architectures, fine-tuning strategies, generalization, and robustness of TSFMs in real-world applications. Code is available at this repository: https://github.com/khanalsubina/msData. Data is available on: \href{https://huggingface.co/datasets/subinak/Open_RAN_Performance_Measurement_Dataset_with_Traffic_and_Mobility_Labels}{Hugging Face}.
With the widespread adoption of AI, machine-to-machine communications are rapidly increasing, reshaping the requirements for optical networks. Recent advances in Gaussian noise modeling for digital coherent transmission have raised expectations for digital-twin-based operation. However, unlike digital twins in wireless communication, which are already well established, significant barriers remain for commercialization in optical networks. This paper discusses the evolving requirements of optical networks in the AI era and proposes a practical Optical Network Digital Twin architecture enabling dynamic and Quality of Transmission aware operation beyond conventional management. Representative use cases, including operator-driven optimization, user-operator collaboration, and multi-operator interconnection, are presented, along with the architectural framework and key challenges toward practical deployment.
Zero Added Loss Multiplexing (ZALM) offers broadband, per channel heralded EPR pairs, with a rich parameter space that allows its performance to be tailored for specific applications. We present a modular ZALM simulator that demonstrates how design choices affect output rate and fidelity. Built in NetSquid with QSI controllers, it exposes 20+ tunable parameters, supports IDEAL and REALISTIC modes, and provides reusable components for Spontaneous Parametric Down Conversion (SPDC) sources, interference, Dense Wavelength Division Multiplexing (DWDM) filtering, fiber delay, active polarization gates, detectors, and lossy fiber. Physics based models capture Hong Ou Mandel (HOM) visibility, insertion loss, detector efficiency, gate errors, and attenuation. Using this tool, we map trade offs among fidelity, link distance, and entangled pairs per use, and show how SPDC bandwidth and DWDM grid spacing steer performance. Using the default configuration settings, average fidelity remains constant at 0.83 but the ebit rate decreases from 0.0175 at the source to 0.0 at 50 km; narrowing the SPDC degeneracy bandwidth increases the ebit rate significantly without affecting fidelity. The simulator enables codesign of source, filtering, and feedforward settings for specific quantum memories and integrates as a building block for end to end quantum network studies.
Resilience in optical networks has traditionally relied on redundancy and pre-planned recovery strategies, both of which assume a certain level of disaster predictability. However, recent environmental changes such as climate shifts, the evolution of communication services, and rising geopolitical risks have increased the unpredictability of disasters, reducing the effectiveness of conventional resilience approaches. To address this unpredictability, this paper introduces the concept of agile resilience, which emphasizes dynamic adaptability across multiple operators and layers. We identify key requirements and challenges, and present enabling technologies for the realization of agile resilience. Using a field-deployed transmission system, we demonstrate rapid system characterization, optical path provisioning, and database migration within six hours. These results validate the effectiveness of the proposed enabling technologies and confirm the feasibility of agile resilience.
Quantum networks rely on high fidelity entangled pairs distributed to nodes, but maintaining their fidelity is challenged by environmental decoherence during storage. Entanglement purification is used to restore fidelity, but the idle periods imposed by the associated classical communication delays counteract this goal by exposing the states to further decoherence. In this work, we analyze the practical viability of entanglement purification protocols (BBPSSW, DEJMPS), under non-instantaneous classical coordination over Internet protocol (IP) communications networks. We present a comprehensive performance evaluation of these protocols in various network conditions for a range of quantum memory technologies. We employ a microscopic Lindblad treatment of the underlying quantum dynamics, and use current-generation metropolitan IP network latency statistics and parameters drawn from quantum memory testbeds. In doing so we identify the regions in which entanglement purification succeeds and fails, delineated by break-even iso-fidelity contours in the phase space. We then determine the total number of entangled pairs required to complete a multi-round purification protocol, and the steady-state throughput of entangled pairs with purified fidelities that exceed application-specific thresholds. This provides latency budgets, memory quality targets, and resource-overhead estimates for deploying purification on current and near-future networks.
Accurate modeling of the gain spectrum in Erbium-Doped Fiber Amplifiers (EDFAs) is essential for optimizing optical network performance, particularly as networks evolve toward multi-vendor solutions. In this work, we propose a generalized few-shot transfer learning architecture based on a Semi-Supervised Self-Normalizing Neural Network (SS-NN) that leverages internal EDFA features - such as VOA input or output power and attenuation, to improve gain spectrum prediction. Our SS-NN model employs a two-phase training strategy comprising unsupervised pre-training with noise-augmented measurements and supervised fine-tuning with a custom weighted MSE loss. Furthermore, we extend the framework with transfer learning (TL) techniques that enable both homogeneous (same-feature space) and heterogeneous (different-feature sets) model adaptation across booster, preamplifier, and ILA EDFAs. To address feature mismatches in heterogeneous TL, we incorporate a covariance matching loss to align second-order feature statistics between source and target domains. Extensive experiments conducted across 26 EDFAs in the COSMOS and Open Ireland testbeds demonstrate that the proposed approach significantly reduces the number of measurements requirements on the system while achieving lower mean absolute errors and improved error distributions compared to benchmark methods.