<p dir="ltr">Federated Learning (FL) is a prime example of a large-scale distributed machine learning framework that has emerged as a result of the exponential growth in data generation and processing capabilities on smart devices. This framework enables the efficient processing and analysis of vast amounts of data, leveraging the collective power of numerous devices to achieve unprecedented scalability and performance. In the FL framework, each end-user device trains a local model using its own data. Through the periodic synchronization of local models, FL achieves a global model that incorporates the insights from all participat- ing devices. This global model can then be used for various applications, such as predictive analytics, recommendation systems, and more.</p><p dir="ltr">Despite its potential, traditional Federated Learning (FL) frameworks face significant hur- dles in real-world applications. These challenges stem from two primary issues: the dynamic nature of data distributions and the efficient utilization of network resources in diverse set- tings. Traditional FL frameworks often rely on the assumption that data distributions remain stationary over time. However, real-world environments are inherently dynamic, with data distributions constantly evolving, which in turn becomes a potential source of <i>temporal</i> het- erogeneity in FL. Another significant challenge in traditional FL frameworks is the efficient use of network resources in heterogeneous settings. Real-world networks consist of devices with varying computational capabilities, communication protocols, and network conditions. Traditional FL frameworks often struggle to adapt to these diverse <i>spatially</i> heterogeneous settings, leading to inefficient use of network resources and increased latency.</p><p dir="ltr">The primary focus of this thesis is to investigate algorithmic frameworks that can miti- gate the challenges posed by <i>temporal</i> and <i>spatial</i> system heterogeneities in FL. One of the significant sources of <i>temporal</i> heterogeneities in FL is owed to the dynamic drifting of client datasets over time, whereas <i>spatial</i> heterogeneities majorly broadly subsume the diverse computational capabilities and network conditions of devices in a network. We introduce two novel FL frameworks: MASTER-FL, which addresses model staleness in the presence of <i>temporally</i> drifting datasets, and Cooperative Edge-Assisted Dynamic Federated Learning CE-FL, which manages both <i>spatial</i> and <i>temporal</i> heterogeneities in extensive hierarchical FL networks. MASTER-FL is specifically designed to ensure that the global model remains accurate and up-to-date even in environments which are characterized by rapidly changing datasets across time. CE-FL, on the other hand, leverages server-side computing capabili- ties, intelligent data offloading, floating aggregation and cooperative learning strategies to manage the diverse computational capabilities and network conditions often associated with modern FL systems.</p>

Le cloud computing a été considéré comme une option pour exécuter des applications de calcul haute performance. Bien que les plateformes traditionnelles de calcul haute performance telles que les grilles et les supercalculateurs offrent un environnement stable du point de vue des défaillances, des performances, et de la taille des ressources, le cloud computing offre des ressources à la demande, généralement avec des performances imprévisibles mais à des coûts financiers abordables. Pour surmonter les limites d’un cloud individuel, plusieurs clouds peuvent être combinés pour former une fédération de clouds, souvent avec des coûts supplémentaires légers pour les utilisateurs. Une fédération de clouds peut aider autant les fournisseurs que les utilisateurs à atteindre leurs objectifs tels la réduction du temps d’exécution, la minimisation des coûts, l’augmentation de la disponibilité, la réduction de la consommation d’énergie, pour ne citer que ceux-Là. Ainsi, la fédération de clouds peut être une solution élégante pour éviter le sur-Approvisionnement, réduisant ainsi les coûts d’exploitation en situation de charge moyenne, et en supprimant des ressources qui, autrement, resteraient inutilisées et gaspilleraient ainsi de énergie. Cependant, la fédération de clouds élargit la gamme des ressources disponibles. En conséquence, pour les utilisateurs, des compétences en cloud computing ou en administration système sont nécessaires, ainsi qu’un temps d’apprentissage considérable pour maîtrises les options disponibles. Dans ce contexte, certaines questions se posent: (a) Quelle ressource du cloud est appropriée pour une application donnée? (b) Comment les utilisateurs peuvent-Ils exécuter leurs applications HPC avec un rendement acceptable et des coûts financiers abordables, sans avoir à reconfigurer les applications pour répondre aux normes et contraintes du cloud ? (c) Comment les non-Spécialistes du cloud peuvent-Ils maximiser l’usage des caractéristiques du cloud, sans être liés au fournisseur du cloud ? et (d) Comment les fournisseurs de cloud peuvent-Ils exploiter la fédération pour réduire la consommation électrique, tout en étant en mesure de fournir un service garantissant les normes de qualité préétablies ? À partir de ces questions, la présente thèse propose une solution de consolidation d’applications pour la fédération de clouds qui garantit le respect des normes de qualité de service. On utilise un système multi-Agents pour négocier la migration des machines virtuelles entre les clouds. En nous basant sur la fédération de clouds, nous avons développé et évalué une approche pour exécuter une énorme application de bioinformatique à coût zéro. En outre, nous avons pu réduire le temps d’exécution de 22,55% par rapport à la meilleure exécution dans un cloud individuel. Cette thèse présente aussi une architecture de cloud baptisée « Excalibur » qui permet l’adaptation automatique des applications standards pour le cloud. Dans l’exécution d’une chaîne de traitements de la génomique, Excalibur a pu parfaitement mettre à l’échelle les applications sur jusqu’à 11 machines virtuelles, ce qui a réduit le temps d’exécution de 63% et le coût de 84% par rapport à la configuration de l’utilisateur. Enfin, cette thèse présente un processus d’ingénierie des lignes de produits (PLE) pour gérer la variabilité de l’infrastructure à la demande du cloud, et une architecture multi-Cloud autonome qui utilise ce processus pour configurer et faire face aux défaillances de manière indépendante. Le processus PLE utilise le modèle étendu de fonction avec des attributs pour décrire les ressources et les sélectionner en fonction des objectifs de l’utilisateur. Les expériences réalisées avec deux fournisseurs de cloud différents montrent qu’en utilisant le modèle proposé, les utilisateurs peuvent exécuter leurs applications dans un environnement de clouds fédérés, sans avoir besoin de connaître les variabilités et contraintes du cloud. / Cloud computing has been seen as an option to execute high performance computing (HPC) applications. While traditional HPC platforms such as grid and supercomputers offer a stable environment in terms of failures, performance, and number of resources, cloud computing offers on-Demand resources generally with unpredictable performance at low financial cost. Furthermore, in cloud environment, failures are part of its normal operation. To overcome the limits of a single cloud, clouds can be combined, forming a cloud federation often with minimal additional costs for the users. A cloud federation can help both cloud providers and cloud users to achieve their goals such as to reduce the execution time, to achieve minimum cost, to increase availability, to reduce power consumption, among others. Hence, cloud federation can be an elegant solution to avoid over provisioning, thus reducing the operational costs in an average load situation, and removing resources that would otherwise remain idle and wasting power consumption, for instance. However, cloud federation increases the range of resources available for the users. As a result, cloud or system administration skills may be demanded from the users, as well as a considerable time to learn about the available options. In this context, some questions arise such as: (a) which cloud resource is appropriate for a given application? (b) how can the users execute their HPC applications with acceptable performance and financial costs, without needing to re-Engineer the applications to fit clouds' constraints? (c) how can non-Cloud specialists maximize the features of the clouds, without being tied to a cloud provider? and (d) how can the cloud providers use the federation to reduce power consumption of the clouds, while still being able to give service-Level agreement (SLA) guarantees to the users? Motivated by these questions, this thesis presents a SLA-Aware application consolidation solution for cloud federation. Using a multi-Agent system (MAS) to negotiate virtual machine (VM) migrations between the clouds, simulation results show that our approach could reduce up to 46% of the power consumption, while trying to meet performance requirements. Using the federation, we developed and evaluated an approach to execute a huge bioinformatics application at zero-Cost. Moreover, we could decrease the execution time in 22.55% over the best single cloud execution. In addition, this thesis presents a cloud architecture called Excalibur to auto-Scale cloud-Unaware application. Executing a genomics workflow, Excalibur could seamlessly scale the applications up to 11 virtual machines, reducing the execution time by 63% and the cost by 84% when compared to a user's configuration. Finally, this thesis presents a product line engineering (PLE) process to handle the variabilities of infrastructure-As-A-Service (IaaS) clouds, and an autonomic multi-Cloud architecture that uses this process to configure and to deal with failures autonomously. The PLE process uses extended feature model (EFM) with attributes to describe the resources and to select them based on users' objectives. Experiments realized with two different cloud providers show that using the proposed model, the users could execute their application in a cloud federation environment, without needing to know the variabilities and constraints of the clouds.

The increasing complexity of telecom networks, induced by the recent development of 5G, is a challenge for detecting faults in the telecom network. In addition to the structural complexity of telecommunication systems, data accessibility has become an issue both in terms of privacy and access cost. We propose a method relying on bi-level Federated Graph Neural Networks to identify anomalies in the telecom network while ensuring reduced communication costs as well as data privacy. Our method considers telecom data as a bi-level graph, where the highest level graph represents the interaction between sites, and each site is further expanded to its software (SW) performance behaviour graph. We developed and compared 4G/5G SW Fault Detection models under 3 settings: (1) Centralized Temporal Graph Neural Networks model: we propose a model to detect anomalies in 4G/5G telecom data. (2) Federated Temporal Graph Neural Networks model: we propose Federated Learning (FL) as a mechanism for privacy-aware training of models for fault detection. (3) Personalized Federated Temporal Graph Neural Networks model: we propose a novel aggregation technique, referred to as FedGraph, leveraging both a graph and the similarities between sites for aggregating the models and proposing models more personalized to each site’s behaviour. We compare the benefits of Federated Learning (FL) models (2) and (3) with centralized training (1) in terms of SW performance data modelling, anomaly detection, and communication cost. The evaluation includes both a scenario with normal functioning sites and a scenario where only a subset of sites exhibit faulty behaviour. The combination of SW execution graphs with GNNs has shown improved modelling performance and minor gains in centralized settings (1). In a normal network context, FL models (2) and (3) perform comparably to centralized training (CL), with slight improvements observed when using the personalized strategy (3). However, in abnormal network scenarios, Federated Learning falls short of achieving comparable detection performance to centralized training. This is due to the unintended learning of abnormal site behaviour, particularly when employing the personalized model (3). These findings highlight the importance of carefully assessing and selecting suitable FL strategies for anomaly detection and model training on telecom network data. / Den ökande komplexiteten i telenäten, som är en följd av den senaste utvecklingen av 5G, är en utmaning när det gäller att upptäcka fel i telenäten. Förutom den strukturella komplexiteten i telekommunikationssystem har datatillgänglighet blivit ett problem både när det gäller integritet och åtkomstkostnader. Vi föreslår en metod som bygger på Federated Graph Neural Networks på två nivåer för att identifiera avvikelser i telenätet och samtidigt säkerställa minskade kommunikationskostnader samt dataintegritet. Vår metod betraktar telekomdata som en graf på två nivåer, där grafen på den högsta nivån representerar interaktionen mellan webbplatser, och varje webbplats utvidgas ytterligare till sin graf för programvarans (SW) prestandabeteende. Vi utvecklade och jämförde 4G/5G SW-feldetekteringsmodeller under 3 inställningar: (1) Central Temporal Graph Neural Networks-modell: vi föreslår en modell för att upptäcka avvikelser i 4G/5G-telekomdata. (2) Federated Temporal Graph Neural Networks-modell: vi föreslår Federated Learning (FL) som en mekanism för integritetsmedveten utbildning av modeller för feldetektering. I motsats till centraliserad inlärning aggregeras lokalt tränade modeller på serversidan och skickas tillbaka till klienterna utan att data läcker ut mellan klienterna och servern, vilket säkerställer integritetsskyddande samarbetsutbildning. (3) Personaliserad Federated Temporal Graph Neural Networks-modell: vi föreslår en ny aggregeringsteknik, kallad FedGraph, som utnyttjar både en graf och likheterna mellan webbplatser för att aggregera modellerna. Vi jämför fördelarna med modellerna Federated Learning (FL) (2) och (3) med centraliserad utbildning (1) när det gäller datamodellering av SW-prestanda, anomalidetektering och kommunikationskostnader. Utvärderingen omfattar både ett scenario med normalt fungerande anläggningar och ett scenario där endast en delmängd av anläggningarna uppvisar felaktigt beteende. Kombinationen av SW-exekveringsgrafer med GNN har visat förbättrad modelleringsprestanda och mindre vinster i centraliserade inställningar (1). I en normal nätverkskontext presterar FL-modellerna (2) och (3) jämförbart med centraliserad träning (CL), med små förbättringar observerade när den personliga strategin används (3). I onormala nätverksscenarier kan Federated Learning dock inte uppnå jämförbar detekteringsprestanda med centraliserad träning. Detta beror på oavsiktlig inlärning av onormalt beteende på webbplatsen, särskilt när man använder den personliga modellen (3). Dessa resultat belyser vikten av att noggrant bedöma och välja lämpliga FL-strategier för anomalidetektering och modellträning på telekomnätdata.

Modern health self-monitoring devices and applications, such as Fitbit and MyFitnessPal, empower users to take concrete actions and set fitness and lifestyle goals based on their recorded trends and statistics. Predicting such trends is beneficial in the road of achieving long-time targets, as the individuals can adjust their diets and habits at any point to guarantee success. The design and implementation of such a system, which also respects user privacy, is the main objective of our work.This application is modelled as a time-series forecasting problem. Given the historical data of users, we aim to predict their eating and lifestyle habits in real-time. We apply the federated learning paradigm to our use-case be- cause of the highly-distributed nature of our data and the privacy concerns of such sensitive recorded information. However, federated learning from het- erogeneous sequences of data can be challenging, as even state-of-the-art ma- chine learning techniques for time-series forecasting can encounter difficulties when learning from very irregular data sequences. Specifically, in the pro- posed healthcare scenario, the machine learning algorithms might fail to cater to users with unique dietary patterns.In this work, we implement a two-step streaming clustering mechanism and group clients that exhibit similar eating and fitness behaviours. The con- ducted experiments prove that learning federatively in this context can achieve very high prediction accuracy, as our predictions are no more than 0.025% far from the ground truth value with respect to the range of each feature. Training separate models for each group of users is shown to be beneficial, especially in terms of the training time, but it is highly dependent on the parameters used for the models and the training process. Our experiments conclude that the configuration used for the general federated model cannot be applied to the clusters of data. However, a decrease in prediction error of more than 45% can be achieved, given the parameters are optimized for each case.Lastly, this work tackles the problem of data privacy by applying state-of- the-art differential privacy techniques. Our empirical study shows that noising the gradients sent to the server is unsuitable for small datasets and cancels out the benefits obtained by prior users’ clustering. On the other hand, noising the training data achieves remarkable results, obtaining a differential privacy level corresponding to an epsilon value of 0.1 with an increase in the observed mean absolute error by a factor of only 0.21. / Moderna apparater och applikationer för självövervakning av hälsa, som Fitbit och MyFitnessPal, ger användarna möjlighet att vidta konkreta åtgärder och sätta fitness- och livsstilsmål baserat på deras dokumenterade trender och statistik. Att förutsäga sådana trender är fördelaktigt för att uppnå långtidsmål, eftersom individerna kan anpassa sina dieter och vanor när som helst för att garantera framgång.Utformningen och implementeringen av ett sådant system, som dessutom respekterar användarnas integritet, är huvudmålet för vårt arbete. Denna appli- kation är modellerad som ett tidsserieprognosproblem. Med avseende på an- vändarnas historiska data är målet att förutsäga deras matvanor och livsstilsva- nor i realtid. Vi tillämpar det federerade inlärningsparadigmet på vårt använd- ningsfall på grund av den mycket distribuerade karaktären av vår data och in- tegritetsproblemen för sådan känslig bokförd information. Federerade lärande från heterogena datasekvenser kan emellertid vara utmanande, eftersom även de modernaste maskininlärningstekniker för tidsserieprognoser kan stöta på svårigheter när de lär sig från mycket oregelbundna datasekvenser. Specifikt i det föreslagna sjukvårdsscenariot kan maskininlärningsalgoritmerna misslyc- kas med att förse användare med unika dietmönster.I detta arbete implementerar vi en tvåstegsströmmande klustermekanism och grupperar användare som uppvisar liknande ät- och fitnessbeteenden. De genomförda experimenten visar att federerade lärande i detta sammanhang kan uppnå mycket hög nogrannhet i förutsägelse, eftersom våra förutsägelser in- te är mer än 0,025% ifrån det sanna värdet med avseende på intervallet för varje funktion. Träning av separata modeller för varje grupp användare visar sig vara fördelaktigt, särskilt gällande träningstiden, men det är mycket be- roende av parametrarna som används för modellerna och träningsprocessen. Våra experiment drar slutsatsen att konfigurationen som används för den all- männa federerade modellen inte kan tillämpas på dataklusterna. Dock kan en minskning av förutsägelsefel på mer än 45% uppnås, givet att parametrarna är optimerade för varje fall.Slutligen hanteras problemet med datasekretess genom att tillämpa bästa tillgängliga differentiell integritetsteknik. Vår empiriska studie visar att adde- ra brus till gradienter som skickas till servern är olämpliga för liten data och avbryter fördelarna med tidigare användares kluster. Däremot, genom att ad- dera brus till träningsdata uppnås anmärkningsvärda resultat. En differentierad integritetsnivå motsvarande ett epsilonvärde på 0,1 med en ökning av det ob- serverade genomsnittliga absoluta felet med en faktor på endast 0,21 erhölls.

This work introduces the available authorization frameworks for the purpose of Single Sign-On functionality within an enterprise, along with the fundamental technicalities. The focus of the work is on SAML 2.0 and OAuth 2.0 frame- works. Following the details related to available protocol flows, supported client profiles and security considerations, the two frameworks are compared in accordance with a set of factors given in a criteria. The report discusses the possibilities provided by a Microsoft Windows based infrastructure, as well as different scenarios and their feasibility in an enterprise environment. The preferred framework, OAuth 2.0, is selected according to the given criteria and the comparative discussions.

Cette thèse s'intéresse à la production et au partage cohérent de connaissances distribuées dans le domaine des sciences de la vie. Malgré l'augmentation constante des capacités de stockage et de calcul des infrastructures informatiques, les approches centralisées pour la gestion de grandes masses de données scientifiques multi-sources deviennent inadaptées pour plusieurs raisons: (i) elles ne garantissent pas l'autonomie des fournisseurs de données qui doivent conserver un certain contrôle sur les données hébergées pour des raisons éthiques et/ou juridiques, (ii) elles ne permettent pas d'envisager le passage à l'échelle des plateformes en sciences computationnelles qui sont la source de productions massives de données scientifiques. Nous nous intéressons, dans le contexte des plateformes collaboratives en sciences de la vie NeuroLOG et VIP, d'une part, aux problématiques de distribution et d'hétérogénéité sous-jacentes au partage de ressources, potentiellement sensibles ; et d'autre part, à la production automatique de connaissances au cours de l'usage de ces plateformes, afin de faciliter l'exploitation de la masse de données produites. Nous nous appuyons sur une approche ontologique pour la modélisation des connaissances et proposons à partir des technologies du web sémantique (i) d'étendre ces plateformes avec des stratégies efficaces, statiques et dynamiques, d'interrogations sémantiques fédérées et (ii) d'étendre leur environnent de traitement de données pour automatiser l'annotation sémantique des résultats d'expérience ``in silico'', à partir de la capture d'informations de provenance à l'exécution et de règles d'inférence spécifiques au domaine. Les résultats de cette thèse, évalués sur l'infrastructure distribuée et contrôlée Grid'5000, apportent des éléments de réponse à trois enjeux majeurs des plateformes collaboratives en sciences computationnelles : (i) un modèle de collaborations sécurisées et une stratégie de contrôle d'accès distribué pour permettre la mise en place d'études multi-centriques dans un environnement compétitif, (ii) des résumés sémantiques d'expérience qui font sens pour l'utilisateur pour faciliter la navigation dans la masse de données produites lors de campagnes expérimentales, et (iii) des stratégies efficaces d'interrogation et de raisonnement fédérés, via les standards du Web Sémantique, pour partager les connaissances capitalisées dans ces plateformes et les ouvrir potentiellement sur le Web de données. Mots-clés: Flots de services et de données scientifiques, Services web sémantiques, Provenance, Web de données, Web sémantique, Fédération de bases de connaissances, Intégration de données distribuées, e-Sciences, e-Santé. / This thesis addresses the issues of coherent distributed knowledge production and sharing in the Life-science area. In spite of the continuously increasing computing and storage capabilities of computing infrastructures, the management of massive scientific data through centralized approaches became inappropriate, for several reasons: (i) they do not guarantee the autonomy property of data providers, constrained, for either ethical or legal concerns, to keep the control over the data they host, (ii) they do not scale and adapt to the massive scientific data produced through e-Science platforms. In the context of the NeuroLOG and VIP Life-science collaborative platforms, we address on one hand, distribution and heterogeneity issues underlying, possibly sensitive, resource sharing ; and on the other hand, automated knowledge production through the usage of these e-Science platforms, to ease the exploitation of the massively produced scientific data. We rely on an ontological approach for knowledge modeling and propose, based on Semantic Web technologies, to (i) extend these platforms with efficient, static and dynamic, transparent federated semantic querying strategies, and (ii) to extend their data processing environment, from both provenance information captured at run-time and domain-specific inference rules, to automate the semantic annotation of ``in silico'' experiment results. The results of this thesis have been evaluated on the Grid'5000 distributed and controlled infrastructure. They contribute to addressing three of the main challenging issues faced in the area of computational science platforms through (i) a model for secured collaborations and a distributed access control strategy allowing for the setup of multi-centric studies while still considering competitive activities, (ii) semantic experiment summaries, meaningful from the end-user perspective, aimed at easing the navigation into massive scientific data resulting from large-scale experimental campaigns, and (iii) efficient distributed querying and reasoning strategies, relying on Semantic Web standards, aimed at sharing capitalized knowledge and providing connectivity towards the Web of Linked Data.

De plus en plus de sources de données liées sont publiées à travers le Web en s'appuyant sur les technologies du Web sémantique, formant ainsi un large réseau de données distribuées. Cependant il est difficile pour les consommateurs de données de profiter de la richesse de ces données, compte tenu de leur distribution, de l'augmentation de leur volume et de l'autonomie des sources de données. Les moteurs fédérateurs de données permettent d'interroger ces sources de données en utilisant des techniques de traitement de requêtes distribuées. Cependant, une mise en œuvre naïve de ces techniques peut générer un nombre considérable de requêtes distantes et de nombreux résultats intermédiaires entraînant ainsi un long temps de traitement des requêtes et des communications réseau coûteuse. Par ailleurs, la sémantique des requêtes distribuées est souvent ignorée. L'expressivité des requêtes, le partitionnement des données et leur réplication sont d'autres défis auxquels doivent faire face les moteurs de requêtes. Pour répondre à ces défis, nous avons d'abord proposé une sémantique des requêtes distribuées compatible avec les standards SPARQL et RDF qui préserve l’expressivité de SPARQL. Nous avons ensuite présenté plusieurs stratégies d'optimisation pour un moteur de requêtes fédérées qui interroge de manière transparente des sources de données distribuées. La performance de ces optimisations est évaluée sur une implémentation d’un moteur de requêtes distribuées SPARQL / Driven by the Semantic Web standards, an increasing number of RDF data sources are published and connected over the Web by data providers, leading to a large distributed linked data network. However, exploiting the wealth of these data sources is very challenging for data consumers considering the data distribution, their volume growth and data sources autonomy. In the Linked Data context, federation engines allow querying these distributed data sources by relying on Distributed Query Processing (DQP) techniques. Nevertheless, a naive implementation of the DQP approach may generate a tremendous number of remote requests towards data sources and numerous intermediate results, thus leading to costly network communications. Furthermore, the distributed query semantics is often overlooked. Query expressiveness, data partitioning, and data replication are other challenges to be taken into account. To address these challenges, we first proposed in this thesis a SPARQL and RDF compliant Distributed Query Processing semantics which preserves the SPARQL language expressiveness. Afterwards, we presented several strategies for a federated query engine that transparently addresses distributed data sources, while managing data partitioning, query results completeness, data replication, and query processing performance. We implemented and evaluated our approach and optimization strategies in a federated query engine to prove their effectiveness.

<p>Over the past decade, distributed representation learning has emerged as a popular alternative to conventional centralized machine learning training. The increasing interest in distributed representation learning, specifically federated learning, can be attributed to its fundamental property that promotes data privacy and communication savings. While conventional ML encourages aggregating data at a central location (e.g., data centers), distributed representation learning advocates keeping data at the source and instead transmitting model parameters across the network. However, since the advent of deep learning, model sizes have become increasingly large often comprising million-billions of parameters, which leads to the problem of communication latency in the learning process. In this thesis, we propose to tackle the problem of communication latency in two different ways: (i) learning private representation of data to enable its sharing, and (ii) reducing the communication latency by minimizing the corresponding long-range communication requirements.</p> <p><br></p> <p>To tackle the former goal, we first start by studying the problem of learning representations that are private yet informative, i.e., providing information about intended ''ally'' targets while hiding sensitive ''adversary'' attributes. We propose Exclusion-Inclusion Generative Adversarial Network (EIGAN), a generalized private representation learning (PRL) architecture that accounts for multiple ally and adversary attributes, unlike existing PRL solutions. We then address the practical constraints of the distributed datasets by developing Distributed EIGAN (D-EIGAN), the first distributed PRL method that learns a private representation at each node without transmitting the source data. We theoretically analyze the behavior of adversaries under the optimal EIGAN and D-EIGAN encoders and the impact of dependencies among ally and adversary tasks on the optimization objective. Our experiments on various datasets demonstrate the advantages of EIGAN in terms of performance, robustness, and scalability. In particular, EIGAN outperforms the previous state-of-the-art by a significant accuracy margin (47% improvement), and D-EIGAN's performance is consistently on par with EIGAN under different network settings.</p> <p><br></p> <p>We next tackle the latter objective - reducing the communication latency - and propose two timescale hybrid federated learning (TT-HF), a semi-decentralized learning architecture that combines the conventional device-to-server communication paradigm for federated learning with device-to-device (D2D) communications for model training. In TT-HF, during each global aggregation interval, devices (i) perform multiple stochastic gradient descent iterations on their individual datasets, and (ii) aperiodically engage in consensus procedure of their model parameters through cooperative, distributed D2D communications within local clusters. With a new general definition of gradient diversity, we formally study the convergence behavior of TT-HF, resulting in new convergence bounds for distributed ML. We leverage our convergence bounds to develop an adaptive control algorithm that tunes the step size, D2D communication rounds, and global aggregation period of TT-HF over time to target a sublinear convergence rate of O(1/t) while minimizing network resource utilization. Our subsequent experiments demonstrate that TT-HF significantly outperforms the current art in federated learning in terms of model accuracy and/or network energy consumption in different scenarios where local device datasets exhibit statistical heterogeneity. Finally, our numerical evaluations demonstrate robustness against outages caused by fading channels, as well favorable performance with non-convex loss functions.</p>

The use of Machine Learning (ML) in business has increased significantly over the past years. Creating high quality and robust models requires a lot of data, which is at times infeasible to obtain. As more people are becoming concerned about their data being misused, data privacy is increasingly strengthened. In 2018, the General Data Protection Regulation (GDPR), was announced within the EU. Models that use either sensitive or personal data to train need to obtain that data in accordance with the regulatory rules, such as GDPR. One other data related issue is that enterprises who wish to collaborate on model building face problems when it requires them to share their private corporate data [36, 38]. In this thesis we will investigate how one might overcome the issue of directly accessing private data when training ML models by employing Federated Learning (FL) [38]. The concept of FL is to allow several silos, i.e. separate parties, to train models with the same objective, using their local data and then with the learned model parameters create a central model. The objective of the central model is to obtain the information learned by the separate models, without ever accessing the raw data itself. This is achieved by averaging the separate models’ weights into the central model. FL thus facilitates opportunities to train a model on large amounts of data from several sources, without the need of having access to the data itself. If one can create a model with this methodology, that is not significantly worse than a model trained on the raw data, then positive effects such as strengthened data privacy, cross-enterprise collaboration and more could be attainable. In this work we have used a financial data set consisting of 25242 equity research reports, provided by Skandinaviska Enskilda Banken (SEB). Each report has a recommendation label, either Buy, Sell or Hold, making this a multi-class classification problem. To evaluate the feasibility of FL we fine-tune the pre-trained Transformer model AlbertForSequenceClassification [37] on the classification task. We create one baseline model using the entire data set and an FL model with different experimental settings, for which the data is distributed both uniformly and non-uniformly. The baseline model is used to benchmark the FL model. Our results indicate that the best FL setting only suffers a small reduction in performance. The baseline model achieves an accuracy of 83.5% compared to 82.8% for the best FL model setting. Further, we find that with an increased number of clients, the performance is worsened. We also found that our FL model was not sensitive to non-uniform data distributions. All in all, we show that FL results in slightly worse generalisation compared to the baseline model, while strongly improving on data privacy, as the central model never accesses the clients’ data. / Företags nyttjande av maskininlärning har de senaste åren ökat signifikant och för att kunna skapa högkvalitativa modeller krävs stora mängder data, vilket kan vara svårt att insamla. Parallellt med detta så ökar också den allmänna förståelsen för hur användandet av data missbrukas, vilket har lätt till ett ökat behov av starkare datasäkerhet. 2018 så trädde General Data Protection Regulation (GDPR) i kraft inom EU, vilken bland annat ställer krav på hur företag skall hantera persondata. Företag med maskininlärningsmodeller som på något sätt använder känslig eller personlig data behöver således ha fått tillgång till denna data i enlighet med de rådande lagar och regler som omfattar datahanteringen. Ytterligare ett datarelaterat problem är då företag önskar att skapa gemensamma maskininlärningsmodeller som skulle kräva att de delar deras bolagsdata [36, 38]. Denna uppsats kommer att undersöka hur Federerad Inlärning [38] kan användas för att skapa maskinlärningsmodeller som överkommer dessa datasäkerhetsrelaterade problem. Federerad Inlärning är en metod för att på ett decentraliserat vis träna maskininlärningsmodeller. Detta omfattar att låta flera aktörer träna en modell var. Varje enskild aktör tränar respektive modell på deras isolerade data och delar sedan endast modellens parametrar till en central modell. På detta vis kan varje enskild modell bidra till den gemensamma modellen utan att den gemensamma modellen någonsin haft tillgång till den faktiska datan. Givet att en modell, skapad med Federerad Inlärning kan uppnå liknande resultat som en modell tränad på rådata, så finns många positiva fördelar så som ökad datasäkerhet och ökade samarbeten mellan företag. Under arbetet har ett dataset, bestående av 25242 finansiella rapporter tillgängliggjort av Skandinaviska Ensilda Banken (SEB) använts. Varje enskild rapport innefattar en rekommendation, antingen Köp, Sälj eller Håll, vilket innebär att vi utför muliklass-klassificering. Med datan tränas den förtränade Transformermodellen AlbertForSequence- Classification [37] på att klassificera rapporterna. En Baseline-modell, vilken har tränats på all rådata och flera Federerade modellkonfigurationer skapades, där bland annat varierande fördelningen av data mellan aktörer från att vara jämnt fördelat till vara ojämnt fördelad. Resultaten visar att den bästa Federerade modellkonfigurationen endast presterar något sämre än Baseline-modellen. Baselinemodellen uppnådde en klassificeringssäkerhet på 83.5% medan den bästa Federerade modellen uppnådde 82.8%. Resultaten visar också att den Federerade modellen inte var känslig mot att variera fördelningen av datamängd mellan aktorerna, samt att med ett ökat antal aktörer så minskar klassificeringssäkerheten. Sammanfattningsvis så visar vi att Federerad Inlärning uppnår nästan lika goda resultat som Baseline-modellen, samtidigt så bidrar metoden till avsevärt bättre datasäkerhet då den centrala modellen aldrig har tillgång till rådata.

Federated learning has gained attention recently for its ability to protect data privacy and distribute computing loads [1]. It overcomes the limitations of traditional machine learning algorithms by allowing computers to train on remote data inputs and build models while keeping participant privacy intact. Traditional machine learning offered a solution by enabling computers to learn patterns and make decisions from data without explicit programming. It opened up new possibilities for automating tasks, recognizing patterns, and making predictions. With the exponential growth of data and advances in computational power, machine learning has become a powerful tool in various domains, driving innovations in fields such as image recognition, natural language processing, autonomous vehicles, and personalized recommendations. traditional machine learning, data is usually transferred to a central server, raising concerns about privacy and security. Centralizing data exposes sensitive information, making it vulnerable to breaches or unauthorized access. Centralized machine learning assumes that all data is available at a central location, which is only sometimes practical or feasible. Some data may be distributed across different locations, owned by different entities, or subject to legal or privacy restrictions. Training a global model in traditional machine learning involves frequent communication between the central server and participating devices. This communication overhead can be substantial, particularly when dealing with large-scale datasets or resource-constrained devices. / Recent studies have uncovered security issues with most of the federated learning models. One common false assumption in the federated learning model is that participants are the attacker and would not use polluted data. This vulnerability enables attackers to train their models using polluted data and then send the polluted updates to the training server for aggregation, potentially poisoning the overall model. In such a setting, it is challenging for an edge server to thoroughly inspect the data used for model training and supervise any edge device. This study evaluates the vulnerabilities present in federated learning and explores various types of attacks that can occur. This paper presents a robust prevention scheme to address these vulnerabilities. The proposed prevention scheme enables federated learning servers to monitor participants actively in real-time and identify infected individuals by introducing an encrypted verification scheme. The paper outlines the protocol design of this prevention scheme and presents experimental results that demonstrate its effectiveness. / Thesis / Doctor of Philosophy (PhD) / federated learning models face significant security challenges and can be vulnerable to attacks. For instance, federated learning models assume participants are not attackers and will not manipulate the data. However, in reality, attackers can compromise the data of remote participants by inserting fake or altering existing data, which can result in polluted training results being sent to the server. For instance, if the sample data is an animal image, attackers can modify it to contaminate the training data. This paper introduces a robust preventive approach to counter data pollution attacks in real-time. It incorporates an encrypted verification scheme into the federated learning model, preventing poisoning attacks without the need for specific attack detection programming. The main contribution of this paper is a mechanism for detection and prevention that allows the training server to supervise real-time training and stop data modifications in each client's storage before and between training rounds. The training server can identify real-time modifications and remove infected remote participants with this scheme.