Pasar al contenido principal

COMPROMISE: PRIVACIDAD EN MACHINE LEARNING COLABORATIVO Y DISTRIBUIDO, SUJETO A REQUISITOS DE RENDIMIENTO Y EFICACIA

PID2020-113795RB-C33

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2020
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDAD DE VIGO

Resultados relacionados

Resultados totales (Incluyendo duplicados): 24
Encontrada(s) 1 página(s)

Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Unleashing the power of decentralized serverless IoT dataflow architecture for the Cloud-to-Edge Continuum: a performance comparison

  • López Escobar, Juan José
  • Díaz Redondo, Rebeca Pilar
  • Gil Castiñeira, Felipe Jose
The advent of new computing and communication trends that link pervasive data sources and consumers, such as Edge Computing, 5G and IIoT, has led to the development of the Cloud-to-Edge Continuum in order to take advantage of the resources available in massive IoT scenarios and to conduct data analysis to leverage intelligence at all levels. This paper outlines the challenging requirements of this novel IoT context and presents an innovative IoT framework to develop dataflow applications for data-centric environments. The proposed design takes advantage of decentralized Pub/Sub communication and serverless nanoservice architecture, using novel technologies such as Zenoh and WebAssembly, respectively, to implement lightweight services along the Cloud-to-Edge infrastructure. We also describe some use cases to illustrate the benefits and concerns of the coming IoT generation, giving a communication performance comparison of Zenoh over brokered MQTT strategies., Ministerio de Universidades | Ref. FPU19/01284, Agencia Estatal de Investigación | Ref. PCI2020-112174, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2020-116329GB-C21, Xunta de Galicia | Ref. GRC-ED431C2022/04 T254, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Contribución de congreso (ConferenceOutput)

Decentralized serverless IoT dataflow architecture for the cloud-to-edge continuum

  • López Escobar, Juan José
  • Gil Castiñeira, Felipe Jose
  • Díaz Redondo, Rebeca Pilar
The advent of new computing and communication trends that link pervasive data sources and consumers, such as Edge Computing, 5G and IIoT, has led to the development of the Cloud-to-Edge Continuum in order to take advantage of the resources available in massive IoT scenarios, and to conduct data analysis to leverage intelligence at all levels. This paper outlines the challenging requirements of this novel IoT context and presents an innovative IoT framework to develop dataflow applications for data-centric environments. The proposed design takes advantage of decentralized Pub/Sub communication and serverless nanoservice architecture, using novel technologies such as Zenoh and WebAssembly, respectively, to implement lightweight services along the Cloud-to-Edge infrastructure. We also describe some use cases to illustrate the benefits and concerns of the coming IoT generation., Ministerio de Universidades | Ref. FPU19/01284, Agencia Estatal de Investigación | Ref. PID2020-113795RBC33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Privacy-aware Berrut Approximated Coded Computing for Federated Learning

  • Martínez Luaña, Xavier
  • Díaz Redondo, Rebeca Pilar
  • Fernández Veiga, Manuel
Federated Learning (FL) is a machine learning framework that enhances privacy compared to normal FL, since model training is done across multiple nodes by exchanging only local parameters instead of raw data. Despite this advantage, FL is still vulnerable to some privacy attacks that have been addressed with techniques such as Differential Privacy (DP), Homomorphic Encryption (HE), or Secure Multi-Party Computation (SMPC). These methods need further assumptions that narrow their range of application, exhibiting problems to work with non-linear functions, to perform large matrix multiplications and having high communication and computational costs to manage semi-honest nodes. In this paper, we propose a solution to guarantee privacy in FL schemes that simultaneously solves these shortcomings. Our proposal is based on the Berrut Approximated Coded Computing (BACC), a technique from the coded computing paradigm, adapted here to provide input privacy to FL in a scalable way. In our Private BACC, controlled randomness is introduced into the rational Berrut approximation of a target function so that the input values are hidden and the output function values can still be computed with good approximation. We provide a theoretical bound to the privacy leakage level of private BACC, and further describe its application for computing arbitrary non-linear functions, particularly distributed matrix multiplications. Our numerical results demonstrate that adding privacy in this form outperforms DP in model quality, since it has negligible cost in the precision. Additionally, it beats HE in terms of computational cost, and requires lower computation and communications costs than with SMPC algorithms. Finally, the proposed method supports distributed matrix multiplications, so it can be integrated into a wide class of machine learning systems., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Instituto Nacional de Ciberseguridad (INCIBE), Xunta de Galicia | Ref. ED431B 2024/41, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Neural collaborative filtering with ontologies for integrated recommendation systems

  • Alaa El-Deen Ahmed, Rana
  • Fernández Veiga, Manuel
  • Gawich, Mariam
Machine learning (ML) and especially deep learning (DL) with neural networks have demonstrated an amazing success in all sorts of AI problems, from computer vision to game playing, from natural language processing to speech and image recognition. In many ways, the approach of ML toward solving a class of problems is fundamentally different than the one followed in classical engineering, or with ontologies. While the latter rely on detailed domain knowledge and almost exhaustive search by means of static inference rules, ML adopts the view of collecting large datasets and processes this massive information through a generic learning algorithm that builds up tentative solutions. Combining the capabilities of ontology-based recommendation and ML-based techniques in a hybrid system is thus a natural and promising method to enhance semantic knowledge with statistical models. This merge could alleviate the burden of creating large, narrowly focused ontologies for complicated domains, by using probabilistic or generative models to enhance the predictions without attempting to provide a semantic support for them. In this paper, we present a novel hybrid recommendation system that blends a single architecture of classical knowledge-driven recommendations arising from a tailored ontology with recommendations generated by a data-driven approach, specifically with classifiers and a neural collaborative filtering. We show that bringing together these knowledge-driven and data-driven worlds provides some measurable improvement, enabling the transfer of semantic information to ML and, in the opposite direction, statistical knowledge to the ontology. Moreover, the novel proposed system enables the extraction of the reasoning recommendation results after updating the standard ontology with the new products and user behaviors, thus capturing the dynamic behavior of the environment of our interest., Ministerio de Ciencia e Innovación | Ref. PID2020-113795RB-C33, Ministerio de Ciencia e Innovación | Ref. RED2018-102585- T
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Capítulo o parte de un libro (BookPart)

¿Cómo enriquecer los recursos digitales musicales para infantil en repositorios gallegos?

  • Casal De La Fuente, Lucia
Introducción Desde el comienzo de la pandemia originada por el Covid-19 parece que el binomio “tecnología” y “mundo formativo y educativo” se ha vuelto indisociable. No obstante, esto es matizable. La tecnología ha tenido siempre su lugar en la esfera educativa en general. No hay más que recurrir a un referente en este campo, Neil Selwyn (2011), para entender a un nivel básico qué es tecnología: “proceso por el cual el ser humano modifica la naturaleza para alcanzar sus necesidades y deseos” (p. 6). Un lápiz, entonces, puede ser considerado tecnología en su sentido más amplio: se crea, por ejemplo, para dar cobertura a la necesidad de dibujar o escribir, y nadie puede afirmar que su presencia en o incorporación a la escuela es una innovación reciente. Por un lado, esta tecnología en su sentido amplio es considerada por Selwyn como tecnología “pre-digital”; mientras que por otro, la tecnología en su sentido más restringido hace referencia a todo lo que tiene que ver con lo digital, o lo que es lo mismo: tecnología digital (hardware, sistemas y aparatos/dispositivos - incluyendo teléfonos móviles y dispositivos audiovisuales como la fotografía digital-; máquinas y videoconsolas; software -programas de simulación, procesadores…-; contenido de la World Wide Web, servicios y aplicaciones, etc. [...], Agencia Estatal de Investigación | Ref. RTI2018-093397-B-100
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Towards efficient compression and communication for prototype-based decentralized learning

  • Fernández Piñeiro, Pablo
  • Fernández Veiga, Manuel
  • Díaz Redondo, Rebeca Pilar
  • Fernández Vilas, Ana
  • González Soto, Martín
In prototype-based federated learning, the exchange of model parameters between clients and the master server is replaced by transmission of prototypes or quantized versions of the data samples to the aggregation server. A fully decentralized deployment of prototype-based learning, without a central aggregator of prototypes, is more robust upon network failures and reacts faster to changes in the statistical distribution of the data, suggesting potential advantages and quick adaptation in dynamic learning tasks, e.g., when the data sources are IoT devices or when data is non-iid. In this paper, we face the challenge of designing an efficient prototype-based decentralized learning network by reducing the overheads in communication and computation. This allows enhancing the scalability of the global system, specially for IoT settings with resource-limited devices. First, we compress the prototype size by applying a clustering algorithm. After that, we filter the prototypes to be disseminate using an information-theoretic measure to share only relevant models or models that provide new knowledge to their neighbors. Then, we define a parallel gossip algorithm to disseminate these models within the learning network. Finally, we define a suitable scheduler able to manage the set of prototypes received to optimize the aggregation phase. In order to validate our proposal we present an analysis of the parallel gossip algorithm regarding the age-of-information (AoI). Our experimental results show the communications load can be substantially reduced without decreasing the convergence rate of the learning algorithm., Universidade de Vigo/CISUG, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Xunta de Galicia | Ref. ED431B 2024/41
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Irregular repetition slotted Aloha with multiuser detection: A density evolution analysis

  • Fernández Veiga, Manuel
  • Sousa Vieira, María Estrella
  • Fernández Vilas, Ana
  • Díaz Redondo, Rebeca Pilar
Irregular repetition slotted Aloha (IRSA) has shown significant advantages as a modern technique for uncoordinated random access with massive number of users due to its capability of achieving theoretically a throughput of 1 packet per slot. When the receiver has also the multi-packet reception of multi-user (MUD) detection property, by applying successive interference cancellation, IRSA also obtains very low packet loss probabilities at low traffic loads, but is unable in general to achieve a normalized throughput close to 1. In this paper, we reconsider the case of IRSA with -MUD receivers and derive the general density evolution equations for the non-asymptotic analysis of the packet loss rate, for arbitrary frame lengths and two operational variants: frame-structured and frameless transmissions. The first one defers transmission attempts until the beginning of the next frame of slots, while the second allows transmission immediately after a packet arrival. Next, using the potential function, we give new capacity bounds on the capacity of the system, showing the threshold arrival rate for zero decoding error probability. Our numerical results illustrate performance in terms of throughput and average delay for -MUD IRSA with finite memory at the receiver, and also with bounded maximum delay, Agencia Estatal de Investigación | Ref. PID2020-113240RB-I00, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

HLF-FSL: a decentralized federated split learning solution for IoT on hyperledger fabric

  • Beis Penedo, Carlos
  • Díaz Redondo, Rebeca Pilar
  • Fernández Vilas, Ana
  • Fernández Veiga, Manuel
  • Troncoso Pastoriza, Francisco Manuel
Collaborative machine learning in sensitive domains demands scalable, privacy-aware and access-controlled solutions for enterprise-grade deployment. Conventional federated learning (FL) relies on a central server, introducing single points of failure and privacy risks, while split learning (SL) partitions models for privacy but scales poorly because of sequential training. We present HLF-FSL, a decentralized architecture that combines federated split learning (FSL) with the permissioned blockchain Hyperledger Fabric (HLF). Chaincode orchestrates split-model execution and peer-to-peer aggregation without a central coordinator, leveraging HLF’s transient fields and Private Data Collections (PDCs) to keep raw data and model activations off-chain and access-controlled. On CIFAR-10, MNIST and ImageNet-Mini, HLF-FSL matches the accuracy of a standard server-coordinated FSL baseline while reducing per-epoch training time versus Ethereum-based baselines. Performance and scalability tests quantify the Fabric coordination overhead via a component-level breakdown of SDK-facing latencies and communication volumes; empirically, this overhead increases wall-clock epoch time while preserving the same accuracy-vs-epoch behavior as a FedSplit Learning baseline., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Xunta de Galicia | Ref. ED431B 2024/41, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Overcoming catastrophic forgetting in tabular data classification: A pseudorehearsal-based approach

  • García Santaclara, Pablo
  • Fernández Castro, Bruno
  • Díaz Redondo, Rebeca Pilar
Continual learning (CL) poses the important challenge of adapting to evolving data distributions without forgetting previously acquired knowledge while consolidating new knowledge. In this paper, we introduce a new methodology, coined the Tabular-data Rehearsal-based Incremental Lifelong Learning framework (TRIL3), designed to address the phenomenon of catastrophic forgetting in an online, task-free setting. TRIL3 utilizes an Incremental Learning Vector Quantization (ILVQ) algorithm as an efficient prototype-based incremental generative model to store and generate synthetic data to preserve knowledge over time, and the Deep Neural Decision Forest (DNDF) algorithm, which was modified to run incrementally to learn supervised classification tasks for tabular data. Based on tests conducted to determine the optimal percentage of synthetic data and comparisons with other available task-free CL proposals, we conclude that TRIL3 outperforms other methods in the literature using only 50% of synthetic data., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Xunta de Galicia | Ref. ED431B 2024/41, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

El rey doliente en la encrucijada de 1847: Don Enrique III, de Ceferino Suárez Bravo, y El gabán del rey, de Gregorio Romero Larrañaga–Eduardo Asquerino, frente a frente

  • Ribao Pereira, Montserrat
En 1847 se publican y estrenan, con pocos meses de diferencia, dos dramas históricos de signo bien diferente: Don Enrique III, de Ceferino Suárez Bravo, y El gabán del rey, de Gregorio Romero Larrañaga y Eduardo Asquerino. Ambos llevan a escena un episodio legendario de la vida del rey Doliente(la venganza del monarca cuyo poder ha sido usurpado por la nobleza y el clero), que se convierte en materia literaria sistemáticamente desde el siglo diecisiete y que el Romanticismo reescribe desde una perspectiva política. En el caso de los dos dramas que me ocupan, de la ideología afín al carlismo de Suárez Bravo y del republicanismo de Asquerino resultan dos productos teatrales de orientación ideológica divergente que explican, contemplados en su conjunto, la complejidad del tiempo en que se escriben., Agencia Estatal de Investigación | Ref. PGC2018-093619-B-I00
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Study of coded ALOHA with multi-user detection under heavy-tailed and correlated arrivals

  • Sousa Vieira, Maria Estrella
  • Fernández Veiga, Manuel
In this paper, we study via simulation the performance of irregular repetition slotted ALOHA under multi-packet detection and different patterns of the load process. On the one hand, we model the arrival process with a version of the M/G/∞ process able to exhibit a correlation structure decaying slowly in time. Given the independence among frames in frame-synchronous coded-slotted ALOHA (CSA), this variation should only take effect on frame-asynchronous CSA. On the other hand, we vary the marginal distribution of the arrival process using discrete versions of the Lognormal and Pareto distributions, with the objective of investigating the influence of the right tail. In this case, both techniques should be affected by the change, albeit to a different degree. Our results confirm these hypotheses and show that these factors must be taken into account when designing and analyzing these systems. In frameless operations, both the shape of the packet arrivals tail distribution and the existence of short-range and long-range correlations strongly impact the packet loss ratio and the average delay. Nevertheless, these effects emerge only weakly in the case of frame-aligned operations, because this enforces the system to introduce a delay in the newly arrived packets (until the beginning of the next frame), and implies that the backlog of accumulated packets is the key quantity for calculating the performance., Ministerio de Ciencia e Innovación | Ref. PID2020-113240RB-I00, Ministerio de Ciencia e Innovación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Decentralized and collaborative machine learning framework for IoT

  • González Soto, Martín
  • Díaz Redondo, Rebeca Pilar
  • Fernández Veiga, Manuel
  • Fernández Castro, Bruno
  • Fernández Vilas, Ana
Decentralized machine learning has recently been proposed as a potential solution to the security issues of the canonical federated learning approach. In this paper, we propose a decentralized and collaborative machine learning framework specially oriented to resource-constrained devices, usual in IoT deployments. With this aim we propose the following construction blocks. First, an incremental learning algorithm based on prototypes that was specifically implemented to work in low-performance computing elements. Second, two random-based protocols to exchange the local models among the computing elements in the network. Finally, two algorithmics approaches for prediction and prototype creation. This proposal was compared to a typical centralized incremental learning approach in terms of accuracy, training time and robustness with very promising results., Axencia Galega de Innovación | Ref. 25/IN606D/2021/2612348, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

In-depth analysis and open challenges of Mist Computing

  • López Escobar, Juan José
  • Díaz Redondo, Rebeca Pilar
  • Gil Castiñeira, Felipe Jose
The advent and consolidation of the Massive Internet of Things (MIoT) comes with a need for new architectures to process the massive amount of generated information. A new approach, Mist Computing, entails a series of changes compared to previous computing paradigms, such as Cloud and Fog Computing, with regard to extremely low latency, local smart processing, high mobility, and massive deployment of heterogeneous devices. Hence, context awareness use cases will be enabled, which will vigorously promote the implementation of advantageous Internet of Things applications. Mist Computing is expected to reach existing fields, such as Industry 4.0, future 6G networks and Big Data problems, and it may be the answer for advanced applications where interaction with the environment is essential and lots of data are managed. Despite the low degree of maturity, it shows plenty of potential for IoT together with Cloud, Fog, and Edge Computing, but it is required to reach a general agreement about its foundations, scope, and fields of action according to the existing early works. In this paper, (i) an extensive review of proposals focused on Mist Computing is done to determine the application fields and network elements that must be developed for certain objectives, besides, (ii) a comparative assessment between Cloud, Fog, Edge, and Mist is completed and (iii) several research challenges are listed for future work. In addition, Mist Computing is the last piece to benefit from the resources of complete network infrastructures in the Fluid Computing paradigm., Ministerio de Ciencia, Innovación y Universidades | Ref. FPU19/01284, Agencia Estatal de Investigación | Ref. PCI2020-112174, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Fondo Europeo de Desarrollo Regional, Xunta de Galicia
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Tesis doctoral (DoctoralThesis)

Design of a new distributed mist, edge and fog architecture in beyond-5G environments for the Internet of Things, Diseño de una nueva arquitectura distribuida “mist”, “edge” y “fog” en entornos “beyond” 5G para el Internet de las Cosas, Deseño dunha nova arquitectura distribuida "mist", "edge" e "fog" en entornos "beyond" 5G para a Internet das Cousas

  • López Escobar, Juan José
We are currently at a time of change in the field of computer networks. Progressively, they have evolved from being a mechanism to interconnect people to turning into the fundamental component of a lot of use cases and areas of business in which the information flows automatically between machines to complete processes, or from the sensors so that algorithms take decisions and apply actions. We speak, therefore, of Internet of Things (IoT). Wireless networks are essential in this evolution, since it is unfeasible to deploy wires to connect millions of devices. Although the technologies of networks of wireless sensors have evolved to satisfy the needs of numerous applications, technologies of cellular communication defined by the 3GPP together with low-power wide-area network (LPWAN) technologies are those that promise a radical change, providing global and continuous coverage for all type of devices. We have seen the first cases of success with technologies like Sigfox and LoRaWAN and mobile communications standardized by the 3GPP (such as NB- IoT, LTE-M, etc.). This thesis aims to create new scalable IoT architectures which are compatible with 5G and beyond environments, regarding resources and complexity. This way, became feasible and explode services that have a positive impact in the society, such as e-health, Industry 4.0, the connected vehicle, etc. There have already been proposals which try to solve issues related to the increase of services based in devices IoT, but it is necessary to go further and propose new distributed architectures which intelligence can be applied in. The objectives of this thesis are the following ones. In the first place, know and analyze the limits of the communication technologies for IoT that are proposed in 3GPP architectures (5G and beyond) or other ones (like Sigfox, LoRa, Weightless, etc.). This analysis will not only limit to the properly said communication, thus it will reach the capacity of the architecture to satisfy end-to-end services (capture information, process it and act). Second, it looks for improvements on the proposed architectures and analyze the captured information by the network in the elements according to computing capacity (the own sensor, the access network, the core of network or the surroundings of the cloud). In third place, look for new mechanisms that allow to lighten the sent information, which is very important in of high data volumes. They will analyze mechanisms like mesh networks and the direct communication combined with centralized communications of the cellular networks or the cell-less architectures (5G, 6G). It will keep as fundamental aim the simplicity in the management. Although the management of the data complicates (analysis in distinct places with data consistency, deliver on time, etc.), the previous configuration of the devices must be minimum or non-existent. The own network must contribute to manage this flexibility. It is an ambitious proposal with high potential. The creation of a scalable, self-configurable and distributed architecture for the analysis of the IoT information with the aim to provide results according to a deadline and using the minimum number of resources will allow to satisfy use cases with high-bandwidth demand, requirements of low latency and high costs., Nos encontramos en un momento de cambio en el ámbito de las redes de comunicación. Progresivamente, han evolucionado desde ser un mecanismo para intercomunicar personas hasta convertirse en el componente fundamental de muchos casos de uso y áreas de negocio en los que la información fluye automáticamente entre máquinas para completar procesos, o desde los sensores para que los algoritmos tomen decisiones y apliquen acciones. Hablamos de Internet of Things (IoT). Las redes inalámbricas son esenciales en esta evolución, ya que es inviable desplegar cables para conectar millones de dispositivos, incluso próximos. Aunque las tecnologías de redes de sensores inalámbricas han evolucionado hasta satisfacer las necesidades de numerosas aplicaciones, las tecnologías de comunicación celular definidas por el 3GPP junto con las redes conocidas como low-power wide-area network (LPWAN) son las que prometen un cambio radical, al proporcionar cobertura global y continua para todo tipo de dispositivos. Hemos visto los primeros casos de éxito con tecnologías como Sigfox y LoRaWAN y posteriormente ya con tecnologías en el entorno de las comunicaciones móviles estandarizadas por el 3GPP (tales como NB-IoT, LTE-M, etc.). Esta tesis busca crear nuevas arquitecturas IoT escalables y compatibles con entornos 5G y beyond, de una manera eficiente en cuanto a recursos y complejidad. Así, hacer viables y explotar servicios que tengan un impacto positivo en la sociedad, como pueden ser la monitorización de la salud, la Industria 4.0, el vehículo conectado, etc. Hemos visto que ya hay propuestas que intentan resolver problemáticas relacionadas con el aumento de los servicios basados en dispositivos IoT, sin embargo, es necesario ir más allá y proponer nuevas arquitecturas distribuidas en las que se pueda aplicar inteligencia. En esta tesis se busca alcanzar varios objetivos. En primer lugar, conocer y analizar los límites existentes de las tecnologías de comunicación para IoT que se proponen en las arquitecturas 3GPP (5G y beyond) u otras (como Sigfox, LoRa, Weightless, etc.). Este análisis no se limitará a la comunicación propiamente dicha, sino que se extenderá a la capacidad de la arquitectura para satisfacer servicios de extremo a extremo (capturar información, procesarla y actuar). En segundo lugar, se busca avanzar sobre las arquitecturas propuestas y analizar hasta qué punto se pueden proponer nuevas arquitecturas distribuidas que implementen el análisis de la información capturada por la red en los elementos con capacidad de cómputo (el propio sensor, la red de acceso, el núcleo de red o el entorno cloud). En tercer lugar, se buscan nuevos mecanismos que permitan aligerar el envío de información, punto muy importante en situaciones en las que se transmite un volumen elevado de datos. Se analizarán mecanismos como las redes malladas y la comunicación directa combinadas con las comunicaciones centralizadas de las redes celulares o las arquitecturas cell-less (5G, 6G). Se mantendrá como objetivo fundamental la simplicidad en la gestión. Es decir, aunque la gestión de los datos se complique (análisis en distintos lugares con consistencia de los datos, cumplir los plazos, etc.), la configuración previa de los dispositivos debe ser mínima o inexistente. La propia red debe aportar los mecanismos que gestionan esta flexibilidad. Se trata de una propuesta ambiciosa con un elevado potencial. La creación de una arquitectura distribuida, autoconfigurable y escalable para el análisis de la información IoT con el objetivo de proporcionar resultados en plazo y consumiendo el mínimo número de recursos permitirá satisfacer casos de uso que de otra forma serían inviables debido a altas demandas de ancho de banda, requisitos de baja latencia y costes elevados., Atopámonos nun momento de cambio no ámbito das redes de comunicación. Progresivamente, evolucionaron desde ser un mecanismo para intercomunicar persoas ata converterse na compoñente fundamental de moitos casos de uso e áreas de negocio nas que a información flúe automaticamente entre máquinas para completar procesos, ou desde os sensores para que os algoritmos tomen decisións e apliquen accións. Falamos, por tanto, do Internet of Things (IoT). As redes sen fíos son esenciais nesta evolución, xa que é inviable despregar cables para conectar millóns de dispositivos, mesmo próximos. Aínda que as tecnoloxías de redes de sensores sen fíos evolucionaron ata satisfacer as necesidades de numerosas aplicacións, as tecnoloxías de comunicación celular definidas polo 3GPP xunto coas redes coñecidas como low-power wide-area network (LPWAN) son as que prometen un cambio radical, ao proporcionar cobertura global e continua para todo tipo de dispositivos. Vimos os primeiros casos de éxito con tecnoloxías como Sigfox e LoRaWAN e posteriormente xa con tecnoloxías na contorna das comunicacións móbiles estandarizadas polo 3GPP (tales como NB- IoT, LTE- M, etc.). Esta tese busca crear novas arquitecturas IoT escalables e compatibles con contornas 5G e beyond, dunha maneira eficiente en canto a recursos e complexidade. Así, facer viables e explotar servizos que teñan un impacto positivo na sociedade, como poden ser a monitoraxe da saúde, a Industria 4.0, o vehículo conectado, etc. Xa hai propostas que tentan resolver problemáticas relacionadas co aumento dos servizos baseados en dispositivos IoT, con todo, é necesario ir máis aló e propoñer novas arquitecturas distribuídas nas que se poida aplicar intelixencia. Nesta tese búscase alcanzar varios obxectivos. En primeiro lugar, coñecer e analizar os límites existentes das tecnoloxías de comunicación para IoT que se propoñen nas arquitecturas 3GPP (5G e beyond) ou outras (como Sigfox, LoRa, Weightless, etc.). Esta análise non se limitará á comunicación propiamente dita, senón que se estenderá á capacidade da arquitectura para satisfacer servizos de extremo a extremo (capturar información, procesala e actuar). En segundo lugar, búscase avanzar sobre as arquitecturas propostas e analizar ata que punto pódense propoñer novas arquitecturas distribuídas que implementen a análise da información capturada pola rede nos elementos con capacidade de cómputo (o propio sensor, a rede de acceso, o núcleo de rede ou a contorna cloud). En terceiro lugar, búscanse novos mecanismos que permitan alixeirar o envío de información, punto moi importante en situacións nas que se transmite un volume elevado de datos. Analizaranse mecanismos como as redes malladas e a comunicación directa combinadas coas comunicacións centralizadas das redes celulares ou as arquitecturas cell-less (5G, 6G). Manterase como obxectivo fundamental a simplicidade na xestión. É dicir, aínda que a xestión dos datos complíquese (análises en distintos lugares con consistencia dos datos, cumprir os prazos, etc.), a configuración previa dos dispositivos debe ser mínima ou inexistente. A propia rede debe achegar os mecanismos que xestionan esta flexibilidade. Trátase dunha proposta ambiciosa cun elevado potencial. A creación dunha arquitectura distribuída, autoconfigurable e escalable para a análise da información IoT co obxectivo de proporcionar resultados en prazo e consumindo o mínimo número de recursos permitirá satisfacer casos de uso que serían inviables doutra forma debido a altas demandas de ancho de banda, requisitos de baixa latencia e custos elevados., Agencia Estatal de Investigación | Ref. PID2020-116329GB-C21, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. FPU19/01284
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Improving recommendations for online retail markets based on ontology evolution

  • Alaa, Rana
  • Gawish, Mariam
  • Fernández Veiga, Manuel
The semantic web is considered to be an extension of the present web. In the semantic web, information is given with well-defined meanings, and thus helps people worldwide to cooperate together and exchange knowledge. The semantic web plays a significant role in describing the contents and services in a machine-readable form. It has been developed based on ontologies, which are deemed the backbone of the semantic web. Ontologies are a key technique with which semantics are annotated, and they provide common comprehensible foundation for resources on the semantic web. The use of semantics and artificial intelligence leads to what is known to be “Smarter Web”, where it will be easy to retrieve what customers want to see on e-commerce platforms, and thus will help users save time and enhance their search for the products they need. The semantic web is used as well as webs 3.0, which helps enhancing systems performance. Previous personalized recommendation methods based on ontologies identify users’ preferences by means of static snapshots of purchase data. However, as the user preferences evolve with time, the one-shot ontology construction is too constrained for capturing individual diverse opinions and users’ preferences evolution over time. This paper will present a novel recommendation system architecture based on ontology evolution, the proposed subsystem architecture for ontology evolution. Furthermore, the paper proposes an ontology building methodology based on a semi-automatic technique as well as development of online retail ontology. Additionally, a recommendation method based on the ontology reasoning is proposed. Based on the proposed method, e-retailers can develop a more convenient product recommendation system to support consumers’ purchase decisions., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

CO-DEFEND: Continuous decentralized federated learning for secure DoH-based threat detection

  • Cajaraville Aboy, Diego
  • Moure Garrido, Marta
  • Beis Penedo, Carlos
  • García Rubio, Carlos
  • Díaz Redondo, Rebeca Pilar
  • Campo, Celeste
  • Fernández Vilas, Ana
  • Fernández Veiga, Manuel
The use of DNS over HTTPS (DoH) tunneling by an attacker to hide malicious activity within encrypted DNS traffic poses a serious threat to network security, as it allows malicious actors to bypass traditional monitoring and intrusion detection systems while evading detection by conventional traffic analysis techniques. Machine Learning (ML) techniques can be used to detect DoH tunnels; however, their effectiveness relies on large datasets containing both benign and malicious traffic. Sharing such datasets across entities is challenging due to privacy concerns. In this work, we propose CO-DEFEND (Continuous Decentralized Federated Learning for Secure DoH-Based Threat Detection), a Decentralized Federated Learning (DFL) framework that enables multiple entities to collaboratively train a classification machine learning model for DoH threat detection while preserving data privacy, enhancing scalability and resilience against single points of failure. The proposed DFL framework provides a realistic implementation for DoH threat detection, enabling multiple entities to train their local models online with incoming DoH flows in real-time batches as they are processed – an approach that fits naturally within modern Internet architectures. This framework adapts four classical machine learning algorithms, Support Vector Machines (SVM), Logistic Regression (LR), Decision Trees (DT), and Random Forest (RF), for federated scenarios and efficient training. In addition, a key methodological feature of CO-DEFEND is the use of DT and RF as model selection rather than aggregation mechanisms, allowing each participant to retain interpretable and locally optimal decision structures while benefiting from collective updates. We compare our proposed method by using the dataset CIRA-CIC-DoHBrw-2020 with existing machine learning approaches, including more computationally complex alternatives such as neural networks, to demonstrate its effectiveness in detecting malicious DoH tunnels while improving scalability and computational efficiency., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C32, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C33, Agencia Estatal de Investigación | Ref. TED2021-130369B-C31, Agencia Estatal de Investigación | Ref. TED2021-130369B-C32, Xunta de Galicia | Ref. ED431B 2024/41, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Resource allocation for dataflow applications in FANETs using anypath routing

  • López Escobar, Juan José
  • Ricardo, Manuel
  • Campos, Rui
  • Gil Castiñeira, Felipe Jose
  • Díaz Redondo, Rebeca Pilar
Management of network resources in advanced IoT applications is a challenging topic due to their distributed nature from the Edge to the Cloud, and the heavy demand of real-time data from many sources to take action in the deployment. FANETs (Flying Ad-hoc Networks) are a clear example of heterogeneous multi-modal use cases, which require strict quality in the network communications, as well as the coordination of the computing capabilities, in order to operate correctly the final service. In this paper, we present a Virtual Network Embedding (VNE) framework designed for the allocation of dataflow applications, composed of nano-services that produce or consume data, in a wireless infrastructure, such as an airborne network. To address the problem, an anypath-based heuristic algorithm that considers the quality demand of the communication between nano-services is proposed, coined as Quality-Revenue Paired Anypath Dataflow VNE (QRPAD-VNE). We also provide a simulation environment for the evaluation of its performance according to the virtual network (VN) request load in the system. Finally, we show the suitability of a multi-parameter framework in conjunction with anypath routing in order to have better performance results that guarantee minimum quality in the wireless communications., Xunta de Galicia | Ref. ED431C 2022/04 T254, Ministerio de Universidades | Ref. FPU19/01284, Agencia Estatal de Investigación | Ref. PCI2020-112174, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2020-116329GB-C21, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

A blockchain solution for decentralized training in machine learning for IoT

  • Beis Penedo, Carlos
  • Troncoso Pastoriza, Francisco Manuel
  • Díaz Redondo, Rebeca Pilar
  • Fernández Vilas, Ana
  • Fernández Veiga, Manuel
  • González Soto, Martín
The rapid growth of Internet of Things (IoT) devices and applications has led to an increased demand for advanced analytics and machine learning techniques capable of handling the challenges associated with data privacy, security, and scalability. Federated learning (FL) and blockchain technologies have emerged as promising approaches to address these challenges by enabling decentralized, secure, and privacy-preserving model training on distributed data sources. In this paper, we present a novel IoT solution that combines the incremental learning vector quantization algorithm (XuILVQ) with Ethereum blockchain technology to facilitate secure and efficient data sharing, model training, and prototype storage in a distributed environment. Our proposed architecture addresses the shortcomings of existing blockchain-based FL solutions by reducing computational and communication overheads while maintaining data privacy and security. We assess the performance of our system through a series of experiments, showing its potential to enhance the accuracy and efficiency of machine learning tasks in IoT settings., Universidade de Vigo/CISUG, Xunta de Galicia | Ref. ED431B 2024/41 (GPC), Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33, Agencia Estatal de Investigación | Ref. PID2023-148716OB-C31, Axencia Galega de Innovación | Ref. 25/IN606D/2021/2612348
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Dental age estimation using deep learning: a comparative survey

  • Mohamed, Essraa Gamal
  • Díaz Redondo, Rebeca Pilar
  • Koura, Abdelrahim
  • EL-Mofty, Mohamed Sherif
  • Kayed, Mohammed
The significance of age estimation arises from its applications in various fields, such as forensics, criminal investigation, and illegal immigration. Due to the increased importance of age estimation, this area of study requires more investigation and development. Several methods for age estimation using biometrics traits, such as the face, teeth, bones, and voice. Among then, teeth are quite convenient since they are resistant and durable and are subject to several changes from childhood to birth that can be used to derive age. In this paper, we summarize the common biometrics traits for age estimation and how this information has been used in previous research studies for age estimation. We have paid special attention to traditional machine learning methods and deep learning approaches used for dental age estimation. Thus, we summarized the advances in convolutional neural network (CNN) models to estimate dental age from radiological images, such as 3D cone-beam computed tomography (CBCT), X-ray, and orthopantomography (OPG) to estimate dental age. Finally, we also point out the main innovations that would potentially increase the performance of age estimation systems., Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Anti-sexism alert system: identification of sexist comments on social media using aI techniques

  • Díaz Redondo, Rebeca Pilar
  • Fernández Vilas, Ana
  • Ramos Merino, Mateo
  • Valladares Rodríguez, Sonia Maria
  • Torres Guijarro, Maria Soledad
  • Hafez, Manar Mohamed
Social relationships in the digital sphere are becoming more usual and frequent, and they constitute a very important aspect for all of us. Violent interactions in this sphere are very frequent, and have serious effects on the victims. Within this global scenario, there is one kind of digital violence that is becoming really worrying: sexism against women. Sexist comments that are publicly posted in social media (newspaper comments, social networks, etc.), usually obtain a lot of attention and become viral, with consequent damage to the persons involved. In this paper, we introduce an anti-sexism alert system, based on natural language processing (NLP) and artificial intelligence (AI), that analyzes any public post, and decides if it could be considered a sexist comment or not. Additionally, this system also works on analyzing all the public comments linked to any multimedia content (piece of news, video, tweet, etc.) and decides, using a color-based system similar to traffic lights, if there is sexism in the global set of posts. We have created a labeled data set in Spanish, since the majority of studies focus on English, to train our system, which offers a very good performance after the validation experiments., Cátedra Feminismos 4.0 Depo-Uvigo, Xunta de Galicia, European Union, Agencia Estatal de Investigación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Ensemble Federated Learning: An approach for collaborative pneumonia diagnosis

  • Mabrouk, Alhassan
  • Díaz Redondo, Rebeca Pilar
  • Abd Elaziz, Mohamed
  • Kayed, Mohammed
Federated learning is a very convenient approach for scenarios where (i) the exchange of data implies privacy concerns and/or (ii) a quick reaction is needed. In smart healthcare systems, both aspects are usually required. In this paper, we work on the first scenario, where preserving privacy is key and, consequently, building a unique and massive medical image data set by fusing different data sets from different medical institutions or research centers (computation nodes) is not an option. We propose an ensemble federated learning (EFL) approach that is based on the following characteristics: First, each computation node works with a different data set (but of the same type). They work locally and apply an ensemble approach combining eight well-known CNN models (densenet169, mobilenetv2, xception, inceptionv3, vgg16, resnet50, densenet121, and resnet152v2) on Chest X-ray images. Second, the best two local models are used to create a local ensemble model that is shared with a central node. Third, the ensemble models are aggregated to obtain a global model, which is shared with the computation nodes to continue with a new iteration. This procedure continues until there are no changes in the best local models. We have performed different experiments to compare our approach with centralized ones (with or without an ensemble approach). The results conclude that our proposal outperforms these ones in Chest X-ray images (achieving an accuracy of 96.63%) and offers very competitive results compared to other proposals in the literature. A source code is provided at the Code Ocean repository: https://codeocean.com/capsule/0530602/tree, Xunta de Galicia | Ref. (Centro de investigación de Galicia accreditation 2019– 2022), Universidade de Vigo/CISUG, Ministerio de Ciencia e Innovación | Ref. PID2020-113795RB-C33
Investigo. Repositorio Institucional de la Universidade de Vigo
Publicación Artículo científico (JournalArticle)

Cybersecurity threat detection based on a UEBA framework using Deep Autoencoders

  • Fuentes, Jose
  • Ortega Fernández, Inés
  • Martinez Villanueva, Nora
  • Sestelo Pérez, Marta
The increasing sophistication of cyberattacks, especially insider and process-related anomalies, poses a major challenge to enterprises, as traditional rule-based or shallow anomaly detection systems often fail to capture complex behavioral patterns. User and Entity Behavior Analytics (UEBA) is a broad branch of data analytics that attempts to build a normal behavioral profile in order to detect anomalous events. Among the techniques used to detect anomalies, deep autoencoders constituted one of the most promising deep learning models on UEBA tasks, allowing explainable detection of security incidents that could lead to the leak of personal data, hijacking of systems, or access to sensitive business information. In this study, we introduced the first implementation of an explainable UEBA-based anomaly detection framework that leveraged deep autoencoders in combination with Doc2Vec, a neural network-based approach that learns the distributed representation of documents, to process both numerical and textual features. Additionally, based on the theoretical foundations of neural networks, we offered a novel proof demonstrating the equivalence of two widely used definitions for fully-connected neural networks. The experimental results demonstrated the proposed framework’s capability to detect real and synthetic anomalies effectively generated from real attack data, showing that the models provided not only correct identification of anomalies but also explainable results that enabled the reconstruction of the possible origin of the anomaly. Compared to existing UEBA and anomaly detection approaches, the novelty of our framework lied in combining explainable multimodal feature processing with formal mathematical guarantees. Our findings suggested that the proposed UEBA framework can be seamlessly integrated into enterprise environments., Agencia Estatal de Investigación | Ref. PID2020-118101GB-I00
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Realistic urban traffic generator using decentralized federated learning for the SUMO simulator

  • Bazán Guillén, Alberto|||0000-0001-8634-6907
  • Beis Penedo, Carlos
  • Cajaraville Aboy, Diego
  • Barbecho Bautista, Pablo
  • Díaz Redondo, Rebeca P.
  • Cruz Llopis, Luis Javier de la|||0000-0003-4171-8310
  • Fernández Vilas, Ana
  • Aguilar Igartua, Mónica|||0000-0002-6518-888X
  • Fernández Veiga, Manuel
Realistic urban traffic simulation is essential for sustainable urban planning and the development of intelligent transportation systems. However, generating high-fidelity, time-varying traffic profiles that accurately reflect real-world conditions, especially in large-scale scenarios, remains a major challenge. Existing methods often suffer from limitations in accuracy, scalability, or raise privacy concerns due to centralized data processing. This work introduces DesRUTGe (Decentralized Realistic Urban Traffic Generator), a novel framework that integrates Deep Reinforcement Learning (DRL) agents with the SUMO simulator to generate realistic 24-hour traffic patterns. A key innovation of DesRUTGe is its use of Decentralized Federated Learning (DFL), wherein each traffic detector and its corresponding urban zone function as an independent learning node. These nodes train local DRL models using minimal historical data and collaboratively refine their performance by exchanging model parameters with selected peers (e.g., geographically adjacent zones), without requiring a central coordinator. Evaluated using real-world data from the city of Barcelona, DesRUTGe outperforms standard SUMO-based tools such as RouteSampler, as well as other centralized learning approaches, by delivering more accurate traffic pattern generation., This work was supported in part by the MICIU/AEI/10.13039/501100011033 (COMPROMISE Project) under Grant PID2020-113795RB-C33; in part by the MCIU/AEI/10.13039/501100011033 (DISCOVERY Project) under Grant PID2023-148716OB-C31; in part by the TRUFFLES: Trusted Framework for Federated Learning Systems, within the Strategic Cybersecurity Projects (INCIBE, Spain); in part by the Recovery, Transformation and Resilience Plan (European Union, Next Generation); in part by the Galician Regional Government under Project ED431B 2024/41 (GPC); in part by the Project Anonymization Technology for AI-Based Analytics of Mobility Data (MOBILYTICS) funded by MCIN/AEI/10.13039/501100011033 under Grant TED2021-129782B-I00; in part by the European Union NextGenerationEU/ PRTR; in part by the Project MultiMO: Datos MultiSectoriales para la Movilidad Obligada (European Union, NextGenerationEU) under Grant TSI-100123-2024-60; in part by the Generalitat de Catalunya under AGAUR Grant 2021 SGR 01413; and in part by the Predoctoral Scholarship for the Training of Research Personnel Associated with the Generación de Conocimiento Project under Grant PRE2021-099830.
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Realistic urban traffic generator using decentralized federated learning for the SUMO simulator

  • Bazán Guillén, Alberto
  • Beis Penedo, Carlos
  • Cajaraville Aboy, Diego
  • Barbecho Bautista, Pablo
  • Díaz Redondo, Rebeca P.
  • Cruz Llopis, Luis Javier de la
  • Fernández Vilas, Ana
  • Aguilar Igartua, Mónica
  • Fernández Veiga, Manuel
Realistic urban traffic simulation is essential for sustainable urban planning and the development of intelligent transportation systems. However, generating high-fidelity, time-varying traffic profiles that accurately reflect real-world conditions, especially in large-scale scenarios, remains a major challenge. Existing methods often suffer from limitations in accuracy, scalability, or raise privacy concerns due to centralized data processing. This work introduces DesRUTGe (Decentralized Realistic Urban Traffic Generator), a novel framework that integrates Deep Reinforcement Learning (DRL) agents with the SUMO simulator to generate realistic 24-hour traffic patterns. A key innovation of DesRUTGe is its use of Decentralized Federated Learning (DFL), wherein each traffic detector and its corresponding urban zone function as an independent learning node. These nodes train local DRL models using minimal historical data and collaboratively refine their performance by exchanging model parameters with selected peers (e.g., geographically adjacent zones), without requiring a central coordinator. Evaluated using real-world data from the city of Barcelona, DesRUTGe outperforms standard SUMO-based tools such as RouteSampler, as well as other centralized learning approaches, by delivering more accurate traffic pattern generation., This work was supported in part by the MICIU/AEI/10.13039/501100011033 (COMPROMISE Project) under Grant PID2020-113795RB-C33; in part by the MCIU/AEI/10.13039/501100011033 (DISCOVERY Project) under Grant PID2023-148716OB-C31; in part by the TRUFFLES: Trusted Framework for Federated Learning Systems, within the Strategic Cybersecurity Projects (INCIBE, Spain); in part by the Recovery, Transformation and Resilience Plan (European Union, Next Generation); in part by the Galician Regional Government under Project ED431B 2024/41 (GPC); in part by the Project Anonymization Technology for AI-Based Analytics of Mobility Data (MOBILYTICS) funded by MCIN/AEI/10.13039/501100011033 under Grant TED2021-129782B-I00; in part by the European Union NextGenerationEU/ PRTR; in part by the Project MultiMO: Datos MultiSectoriales para la Movilidad Obligada (European Union, NextGenerationEU) under Grant TSI-100123-2024-60; in part by the Generalitat de Catalunya under AGAUR Grant 2021 SGR 01413; and in part by the Predoctoral Scholarship for the Training of Research Personnel Associated with the Generación de Conocimiento Project under Grant PRE2021-099830., Postprint (published version)