arXiv subjects

Marco Miozzo

Publications and source records attributed to Marco Miozzo.

16 recordsLinked to original sources

arXiv2026

Degradation of Feature Space in Continual Learning

Centralized training is the standard paradigm in deep learning, enabling models to learn from a unified dataset in a single location. In such setup, isotropic feature distributions naturally arise as a mean to support well-structured and generalizable representations. In contrast, continual learning operates on streaming and non-stationary data, and trains models incrementally, inherently facing the well-known plasticity-stability dilemma. In such settings, learning dynamics tends to yield increasingly anisotropic feature space. This arises a fundamental question: should isotropy be enforced to achieve a better balance between stability and plasticity, and thereby mitigate catastrophic forgetting? In this paper, we investigate whether promoting feature-space isotropy can enhance representation quality in continual learning. Through experiments using contrastive continual learning techniques on CIFAR-10 and CIFAR-100 data, we find that isotropic regularization fails to improve, and can in fact degrade, model accuracy in continual settings. Our results highlight essential differences in feature geometry between centralized and continual learning, suggesting that isotropy, while beneficial in centralized setups, may not constitute an appropriate inductive bias for non-stationary learning scenarios.

cs.LG↗

arXiv2025

Self-Supervised Learning at the Edge: The Cost of Labeling

Contrastive learning (CL) has recently emerged as an alternative to traditional supervised machine learning solutions by enabling rich representations from unstructured and unlabeled data. However, CL and, more broadly, self-supervised learning (SSL) methods often demand a large amount of data and computational resources, posing challenges for deployment on resource-constrained edge devices. In this work, we explore the feasibility and efficiency of SSL techniques for edge-based learning, focusing on trade-offs between model performance and energy efficiency. In particular, we analyze how different SSL techniques adapt to limited computational, data, and energy budgets, evaluating their effectiveness in learning robust representations under resource-constrained settings. Moreover, we also consider the energy costs involved in labeling data and assess how semi-supervised learning may assist in reducing the overall energy consumed to train CL models. Through extensive experiments, we demonstrate that tailored SSL strategies can achieve competitive performance while reducing resource consumption by up to 4X, underscoring their potential for energy-efficient learning at the edge.

cs.LG↗

arXiv2025

Energy Minimization for Participatory Federated Learning in IoT Analyzed via Game Theory

The Internet of Things requires intelligent decision making in many scenarios. To this end, resources available at the individual nodes for sensing or computing, or both, can be leveraged. This results in approaches known as participatory sensing and federated learning, respectively. We investigate the simultaneous implementation of both, through a distributed approach based on empowering local nodes with game theoretic decision making. A global objective of energy minimization is combined with the individual node's optimization of local expenditure for sensing and transmitting data over multiple learning rounds. We present extensive evaluations of this technique, based on both a theoretical framework and experiments in a simulated network scenario with real data. Such a distributed approach can reach a desired level of accuracy for federated learning without a centralized supervision of the data collector. However, depending on the weight attributed to the local costs of the single node, it may also result in a significantly high Price of Anarchy (from 1.28 onwards). Thus, we argue for the need of incentive mechanisms, possibly based on Age of Information of the single nodes.

cs.LG↗

arXiv2024

Federated Learning in Mobile Networks: A Comprehensive Case Study on Traffic Forecasting

The increasing demand for efficient resource allocation in mobile networks has catalyzed the exploration of innovative solutions that could enhance the task of real-time cellular traffic prediction. Under these circumstances, federated learning (FL) stands out as a distributed and privacy-preserving solution to foster collaboration among different sites, thus enabling responsive near-the-edge solutions. In this paper, we comprehensively study the potential benefits of FL in telecommunications through a case study on federated traffic forecasting using real-world data from base stations (BSs) in Barcelona (Spain). Our study encompasses relevant aspects within the federated experience, including model aggregation techniques, outlier management, the impact of individual clients, personalized learning, and the integration of exogenous sources of data. The performed evaluation is based on both prediction accuracy and sustainability, thus showcasing the environmental impact of employed FL algorithms in various settings. The findings from our study highlight FL as a promising and robust solution for mobile traffic prediction, emphasizing its twin merits as a privacy-conscious and environmentally sustainable approach, while also demonstrating its capability to overcome data heterogeneity and ensure high-quality predictions, marking a significant stride towards its integration in mobile traffic management systems.

cs.LG↗

arXiv2024

Energy-Aware Decentralized Learning with Intermittent Model Training

Decentralized learning (DL) offers a powerful framework where nodes collaboratively train models without sharing raw data and without the coordination of a central server. In the iterative rounds of DL, models are trained locally, shared with neighbors in the topology, and aggregated with other models received from neighbors. Sharing and merging models contribute to convergence towards a consensus model that generalizes better across the collective data captured at training time. In addition, the energy consumption while sharing and merging model parameters is negligible compared to the energy spent during the training phase. Leveraging this fact, we present SkipTrain, a novel DL algorithm, which minimizes energy consumption in decentralized learning by strategically skipping some training rounds and substituting them with synchronization rounds. These training-silent periods, besides saving energy, also allow models to better mix and finally produce models with superior accuracy than typical DL algorithms that train at every round. Our empirical evaluations with 256 nodes demonstrate that SkipTrain reduces energy consumption by 50% and increases model accuracy by up to 12% compared to D-PSGD, the conventional DL algorithm.

cs.LG↗

arXiv2023

Mobile Traffic Prediction at the Edge Through Distributed and Deep Transfer Learning

Traffic prediction represents one of the crucial tasks for smartly optimizing the mobile network. Recently, Artificial Intelligence (AI) has attracted attention to solve this problem thanks to its ability in cognizing the state of the mobile network and make intelligent decisions. Research on this topic has concentrated on making predictions in a centralized fashion, i.e., by collecting data from the different network elements and process them in a cloud center. This translates into inefficiencies due to the large amount of data transmissions and computations required, leading to high energy consumption. In this work, we investigate a fully decentralized AI solution for mobile traffic prediction that allows data to be kept locally, reducing energy consumption through collaboration among the base station sites. To do so, we propose a novel prediction framework based on edge computing and Deep Transfer Learning (DTL) techniques, using datasets obtained at the edge through a large measurement campaign. Two main Deep Learning architectures are designed based on Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) and tested under different training conditions. Simulation results show that the CNN architectures outperform the RNNs in accuracy and consume less energy. In both scenarios, DTL contributes to an accuracy enhancement in 85% of the examined cases compared to their stand-alone counterparts. Additionally, DTL significantly reduces computational complexity and energy consumption during training, resulting in a reduction of the energy footprint by 60% for CNNs and 90% for RNNs. Finally, two cutting-edge eXplainable Artificial Intelligence techniques are employed to interpret the derived learning models.

cs.DC↗

arXiv2023

Towards Energy-Aware Federated Traffic Prediction for Cellular Networks

Cellular traffic prediction is a crucial activity for optimizing networks in fifth-generation (5G) networks and beyond, as accurate forecasting is essential for intelligent network design, resource allocation and anomaly mitigation. Although machine learning (ML) is a promising approach to effectively predict network traffic, the centralization of massive data in a single data center raises issues regarding confidentiality, privacy and data transfer demands. To address these challenges, federated learning (FL) emerges as an appealing ML training framework which offers high accurate predictions through parallel distributed computations. However, the environmental impact of these methods is often overlooked, which calls into question their sustainability. In this paper, we address the trade-off between accuracy and energy consumption in FL by proposing a novel sustainability indicator that allows assessing the feasibility of ML models. Then, we comprehensively evaluate state-of-the-art deep learning (DL) architectures in a federated scenario using real-world measurements from base station (BS) sites in the area of Barcelona, Spain. Our findings indicate that larger ML models achieve marginally improved performance but have a significant environmental impact in terms of carbon footprint, which make them impractical for real-world applications.

cs.LG↗

arXiv2022

The Cost of Training Machine Learning Models over Distributed Data Sources

Federated learning is one of the most appealing alternatives to the standard centralized learning paradigm, allowing a heterogeneous set of devices to train a machine learning model without sharing their raw data. However, it requires a central server to coordinate the learning process, thus introducing potential scalability and security issues. In the literature, server-less federated learning approaches like gossip federated learning and blockchain-enabled federated learning have been proposed to mitigate these issues. In this work, we propose a complete overview of these three techniques proposing a comparison according to an integral set of performance indicators, including model accuracy, time complexity, communication overhead, convergence time, and energy consumption. An extensive simulation campaign permits to draw a quantitative analysis considering both feedforward and convolutional neural network models. Results show that gossip federated learning and standard federated solution are able to reach a similar level of accuracy, and their energy consumption is influenced by the machine learning model adopted, the software library, and the hardware used. Differently, blockchain-enabled federated learning represents a viable solution for implementing decentralized learning with a higher level of security, at the cost of an extra energy usage and data sharing. Finally, we identify open issues on the two decentralized federated learning implementations and provide insights on potential extensions and possible research directions in this new research field.

cs.LG↗

arXiv2020

Distributed Deep Reinforcement Learning for Functional Split Control in Energy Harvesting Virtualized Small Cells

To meet the growing quest for enhanced network capacity, mobile network operators (MNOs) are deploying dense infrastructures of small cells. This, in turn, increases the power consumption of mobile networks, thus impacting the environment. As a result, we have seen a recent trend of powering mobile networks with harvested ambient energy to achieve both environmental and cost benefits. In this paper, we consider a network of virtualized small cells (vSCs) powered by energy harvesters and equipped with rechargeable batteries, which can opportunistically offload baseband (BB) functions to a grid-connected edge server depending on their energy availability. We formulate the corresponding grid energy and traffic drop rate minimization problem, and propose a distributed deep reinforcement learning (DDRL) solution. Coordination among vSCs is enabled via the exchange of battery state information. The evaluation of the network performance in terms of grid energy consumption and traffic drop rate confirms that enabling coordination among the vSCs via knowledge exchange achieves a performance close to the optimal. Numerical results also confirm that the proposed DDRL solution provides higher network performance, better adaptation to the changing environment, and higher cost savings with respect to a tabular multi-agent reinforcement learning (MRL) solution used as a benchmark.

cs.NI↗

arXiv2020

Recurrent Neural Networks for Handover Management in Next-Generation Self-Organized Networks

In this paper, we discuss a handover management scheme for Next Generation Self-Organized Networks. We propose to extract experience from full protocol stack data, to make smart handover decisions in a multi-cell scenario, where users move and are challenged by deep zones of an outage. Traditional handover schemes have the drawback of taking into account only the signal strength from the serving, and the target cell, before the handover. However, we believe that the expected Quality of Experience (QoE) resulting from the decision of target cell to handover to, should be the driving principle of the handover decision. In particular, we propose two models based on multi-layer many-to-one LSTM architecture, and a multi-layer LSTM AutoEncoder (AE) in conjunction with a MultiLayer Perceptron (MLP) neural network. We show that using experience extracted from data, we can improve the number of users finalizing the download by 18%, and we can reduce the time to download, with respect to a standard event-based handover benchmark scheme. Moreover, for the sake of generalization, we test the LSTM Autoencoder in a different scenario, where it maintains its performance improvements with a slight degradation, compared to the original scenario.

cs.NI↗

arXiv2019

Dynamic Control of Functional Splits for Energy Harvesting Virtual Small Cells: a Distributed Reinforcement Learning Approach

In this paper, we propose a network scenario where the baseband processes of the virtual small cells powered solely by energy harvesters and batteries can be opportunistically executed in a grid-connected edge computing server, co-located at the macro base station site. We state the corresponding energy minimization problem and propose multi-agent Reinforcement Learning (RL) to solve it. Distributed Fuzzy Q-Learning and Q-Learning on-line algorithms are tailored for our purposes. Coordination among the multiple agents is achieved by broadcasting system level information to the independent learners. The evaluation of the network performance confirms that coordination via broadcasting may achieve higher system level gains than un-coordinated solutions and cumulative rewards closer to the off-line bounds. Finally, our analysis permits to evaluate the benefits of continuous state/action representation for the learning algorithms in terms of faster convergence, higher cumulative reward and more adaptation to changing environments.

eess.SY↗

arXiv2018

Optimal Placement of Baseband Functions for Energy Harvesting Virtual Small Cells

Flexible functional split in Cloud Radio Access Network (CRAN) greatly overcomes fronthaul capacity and latency challenges. In such architecture, part of the baseband processing is done locally and the remaining is done remotely in the central cloud. On the other hand, Energy Harvesting (EH) technologies are increasingly adopted due to sustainability and economic advantages. Power consumption due to baseband processing has a huge share in the total power consumption breakdown of smaller base stations. Given that such base stations are powered by EH, in addition to QoS constraints, energy availability also conditions the decision on where to place each baseband function in the system. This work focuses on determining the performance bounds of an optimal placement of baseband functional split option in virtualized small cells that are solely powered by EH. The work applies Dynamic Programming (DP), in particular, Shortest Path search is used to determine the optimal functional split option considering traffic QoS requirements and available energy budget.

cs.NI↗

arXiv2018

End-to-end 5G services via an SDN/NFV-based multi-tenant network and cloud testbed

5G has a main requirement of highly flexible, ultralow latency and ultra-high bandwidth virtualized infrastructure in order to deliver end-to-end services. This requirement can be met by efficiently integrating all network segments (radio access, aggregation and core) with heterogeneous wireless and optical technologies (5G, mmWave, LTE/LTE-A, Wi-Fi, Ethernet, MPLS, WDM, software-defined optical transmission, etc.), and massive computing and storage cloud services (offered in edge/core data centers). This paper introduces the preliminary architecture aiming at integrating three consolidated and standalone experimental infrastructures at CTTC, in order to deploy the required end-to-end top-to-bottom converged infrastructure pointed out above for testing and developing advanced 5G services.

cs.NI↗

arXiv2018

Energy sustainable paradigms and methods for future mobile networks: A survey

In this survey, we discuss the role of energy in the design of future mobile networks and, in particular, we advocate and elaborate on the use of energy harvesting (EH) hardware as a means to decrease the environmental footprint of 5G technology. To take full advantage of the harvested (renewable) energy, while still meeting the quality of service required by dense 5G deployments, suitable management techniques are here reviewed, highlighting the open issues that are still to be solved to provide eco-friendly and cost-effective mobile architectures. Several solutions have recently been proposed to tackle capacity, coverage and efficiency problems, including: C-RAN, Software Defined Networking (SDN) and fog computing, among others. However, these are not explicitly tailored to increase the energy efficiency of networks featuring renewable energy sources, and have the following limitations: (i) their energy savings are in many cases still insufficient and (ii) they do not consider network elements possessing energy harvesting capabilities. In this paper, we systematically review existing energy sustainable paradigms and methods to address points (i) and (ii), discussing how these can be exploited to obtain highly efficient, energy self-sufficient and high capacity networks. Several open issues have emerged from our review, ranging from the need for accurate energy, transmission and consumption models, to the lack of accurate data traffic profiles, to the use of power transfer, energy cooperation and energy trading techniques. These challenges are here discussed along with some research directions to follow for achieving sustainable 5G systems.

cs.NI↗

arXiv2016

Simulating LTE and Wi-Fi Coexistence in Unlicensed Spectrum with ns-3

The use of unlicensed spectrum for future LTE systems raises concerns about its impact on co-located Wi-Fi networks. We report herein on extensions to a popular network simulator, known as ns-3, to model such coexistence. To our knowledge, this simulator is novel in two important ways. First, it is presently the only freely available simulator for coexistence studies, and as open source software, it offers new opportunities for reproducible research and collaborative model development. Second, we are not aware of other simulators with similar wireless models that allow full system studies to be conducted, such as performance evaluation of upper-layer protocols in a coexistence environment. We illustrate the value of a system simulator by describing a simulation campaign for the indoor scenario designed by 3GPP RAN1 using new ns-3 models for Release 13 LTE Listen Before Talk (LBT) techniques, intended for fair coexistence with Wi-Fi. LBT models have been widely simulated at the link and physical layer by many companies in 3GPP RAN1. We first show that we are able to tune the scenario to reproduce similar performance metrics to those reported in the literature. We then extend the simulations by adding TCP to the file transfers and explore the different behaviour observed. Our initial study reveals that there remain many open questions about how LBT should be implemented to achieve good coexistence properties, and that a broader scope of implementation issues, traffic models, and fuller system models should be studied further before drawing final conclusions on the LBT coexistence performance.

cs.NI↗

arXiv2013

SolarStat: Modeling Photovoltaic Sources through Stochastic Markov Processes

In this paper, we present a methodology and a tool to derive simple but yet accurate stochastic Markov processes for the description of the energy scavenged by outdoor solar sources. In particular, we target photovoltaic panels with small form factors, as those exploited by embedded communication devices such as wireless sensor nodes or, concerning modern cellular system technology, by small-cells. Our models are especially useful for the theoretical investigation and the simulation of energetically self-sufficient communication systems including these devices. The Markov models that we derive in this paper are obtained from extensive solar radiation databases, that are widely available online. Basically, from hourly radiance patterns, we derive the corresponding amount of energy (current and voltage) that is accumulated over time, and we finally use it to represent the scavenged energy in terms of its relevant statistics. Toward this end, two clustering approaches for the raw radiance data are described and the resulting Markov models are compared against the empirical distributions. Our results indicate that Markov models with just two states provide a rough characterization of the real data traces. While these could be sufficiently accurate for certain applications, slightly increasing the number of states to, e.g., eight, allows the representation of the real energy inflow process with an excellent level of accuracy in terms of first and second order statistics. Our tool has been developed using Matlab(TM) and is available under the GPL license at[1].

cs.OH↗