JDJérémie Decouchant

Jérémie Decouchant

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jérémie Decouchant has a total of 84 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Automated Network-Level Fault Injection Testing of Microservice Architectures

    by Delano Flipse, Hakan Simsek, Jérémie Decouchant, Burcu Kulahcioglu Ozkan

  • Collusions and Privacy in Rational-Resilient Gossip

    by JérÉmIe Decouchant

    Coalitions et respect de la vie privée dans les protocoles de dissémination aléatoire de contenus tolérant les comportements égoïstes: Les protocoles de dissémination de contenus randomisés sont une alternative bon marché et pouvant monter en charge aux systèmes centralisés. Cependant, il est bien connu que ces protocoles souffrent en présence de comportements individualistes, i.e., de participants qui cherchent à recevoir un contenu sans contribuer en retour à sa propagation. Alors que le problème des participants égoïstes a été bien étudié dans la littérature, les coalitions de participants égoïstes ont été laissés de côté. De plus, les manières actuelles permettant de limiter ou tolérer ces comportements exigent des noeuds qu'ils enregistrent leurs interactions, et rendent public leur contenu, ce qui peut dévoiler des informations gênantes. De nos jours, il y a consensus autour du besoin de renforcer les possibilités de contrôle des usagers de systèmes informatiques sur leurs données personnelles. Cependant, en l'état de nos connaissances, il n'existe pas de protocole qui évite de divulguer des informations personnelles sur les utilisateurs tout en limitant l'impact des comportements individualistes.Cette thèse apporte deux contributions.Tout d'abord, nous présentons AcTinG, un protocole qui empêche les coalitions de noeuds individualistes dans les systèmes pair-à-pair de dissémination de contenus, tout en garantissant une absence de faux-positifs dans le processus de détection de fautes. Les utilisateurs de AcTinG enregistrent leurs interactions dans des enregistrements sécurisés, et se vérifient les uns les autres grâce à une procédure d'audit non prédictible, mais vérifiable a posteriori. Ce protocole est un équilibre de Nash par construction. Une évaluation de performance montre qu'AcTinG est capable de fournir les messages à tous les noeuds malgré la présence de coalitions, et présente des propriétés de passage à l'échelle similaires aux protocoles classiques de dissémination aléatoire.Ensuite, nous décrivons PAG, le premier protocole qui évite de dévoiler des informations sur les usagers tout en les contrôlant afin d'éviter les comportements égoïstes. PAG se base sur une architecture de surveillance formée par les participants, ainsi que sur des procédures de chiffrement homomorphiques. L'évaluation théorique de ce protocole montre qu'obtenir le détail des interactions des noeuds est difficile, même en cas d'attaques collectives. Nous évaluons ce protocole en terme de protection de l'intimité des interactions et en terme de performance en utilisant un déploiement effectué sur un cluster de machines, ainsi que des simulations qui impliquent jusqu'à un million de participants, et enfin en utilisant des preuves théoriques. Ce protocole a un surcoût en bande-passante inférieur aux protocoles de communications anonymes existants, et est raisonnable en terme de coût cryptographique.

  • Open Access

    Dynamic Topology Optimization for Non-IID Data in Decentralized Learning

    by Bart Cox, Antreas Ioannou, Jérémie Decouchant

    Decentralized learning (DL) enables a set of nodes to train a model collaboratively without central coordination, offering benefits for privacy and scalability. However, DL struggles to train a high accuracy model when the data distribution is non-independent and identically distributed (non-IID) and when the communication topology is static. To address these issues, we propose Morph, a topology optimization algorithm for DL. In Morph, nodes adaptively choose peers for model exchange based on maximum model dissimilarity. Morph maintains a fixed in-degree while dynamically reshaping the communication graph through gossip-based peer discovery and diversity-driven neighbor selection, thereby improving robustness to data heterogeneity. Experiments on CIFAR-10 and FEMNIST with up to 100 nodes show that Morph consistently outperforms static and epidemic baselines, while closely tracking the fully connected upper bound. On CIFAR-10, Morph achieves a relative improvement of 1.12x in test accuracy compared to the state-of-the-art baselines. On FEMNIST, Morph achieves an accuracy that is 1.08x higher than Epidemic Learning. Similar trends hold for 50 node deployments, where Morph narrows the gap to the fully connected upper bound within 0.5 percentage points on CIFAR-10. These results demonstrate that Morph achieves higher final accuracy, faster convergence, and more stable learning as quantified by lower inter-node variance, while requiring fewer communication rounds than baselines and no global knowledge.

  • CCBNet : Confidential Collaborative Bayesian Networks Inference

    by Abele Mălan, Thiago Guzella, Jérémie Decouchant, Lydia Chen


  • Open Access

    Balancing Fairness and Performance in Multi-User Spark Workloads with Dynamic Scheduling

    by Dāvis Kažemaks, Laurens Versluis, Burcu Kulahcioglu Ozkan, Jérémie Decouchant

  • Open Access

    NEMO: Faster Parallel Execution for Highly Contended Blockchain Workloads (Full version)

    by François Ezard, Can Umut Ileri, Jérémie Decouchant

    Following the design of more efficient blockchain consensus algorithms, the execution layer has emerged as the new performance bottleneck of blockchains, especially under high contention. Current parallel execution frameworks either rely on optimistic concurrency control (OCC) or on pessimistic concurrency control (PCC), both of which see their performance decrease when workloads are highly contended, albeit for different reasons. In this work, we present NEMO, a new blockchain execution engine that combines OCC with the object data model to address this challenge. NEMO introduces four core innovations: (i) a greedy commit rule for transactions using only owned objects; (ii) refined handling of dependencies to reduce re-executions; (iii) the use of incomplete but statically derivable read/write hints to guide execution; and (iv) a priority-based scheduler that favors transactions that unblock others. Through simulated execution experiments, we demonstrate that NEMO significantly reduces redundant computation and achieves higher throughput than representative approaches. For example, with 16 workers NEMO's throughput is up to 42% higher than the one of Block-STM, the state-of-the-art OCC approach, and 61% higher than the pessimistic concurrency control baseline used.

  • Open Access

    Go With The Flow: Churn-Tolerant Decentralized Training of Large Language Models

    by Nikolay Blagoev, Bart Cox, Jérémie Decouchant, Lydia Y. Chen

    Motivated by the emergence of large language models (LLMs) and the importance of democratizing their training, we propose GWTF, the first crash tolerant practical decentralized training framework for LLMs. Differently from existing distributed and federated training frameworks, GWTF enables the efficient collaborative training of a LLM on heterogeneous clients that volunteer their resources. In addition, GWTF addresses node churn, i.e., clients joining or leaving the system at any time, and network instabilities, i.e., network links becoming unstable or unreliable. The core of GWTF is a novel decentralized flow algorithm that finds the most effective routing that maximizes the number of microbatches trained with the lowest possible delay. We extensively evaluate GWTF on GPT-like and LLaMa-like models and compare it against the prior art. Our results indicate that GWTF reduces the training time by up to 45% in realistic and challenging scenarios that involve heterogeneous client nodes distributed over 10 different geographic locations with a high node churn rate.

  • Share Secrets for Privacy: Confidential Forecasting with Vertical Federated Learning

    by Aditya Shankar, Jérémie Decouchant, Dimitra Gkorou, Rihan Hai, Lydia Chen


  • Open Access

    MUDGUARD: Taming Malicious Majorities in Federated Learning using Privacy-preserving Byzantine-robust Clustering

    by Rui Wang, Xingkai Wang, Huanhuan Chen, Jérémie Decouchant, Stjepan Picek, Nikolaos Laoutaris, Kaitai Liang

    Byzantine-robust Federated Learning (FL) aims to counter malicious clients and train an accurate global model while maintaining an extremely low attack success rate. Most existing systems, however, are only robust when most of the clients are honest. FLTrust (NDSS '21) and Zeno++ (ICML '20) do not make such an honest majority assumption but can only be applied to scenarios where the server is provided with an auxiliary dataset used to filter malicious updates. FLAME (USENIX '22) and EIFFeL (CCS '22) maintain the semi-honest majority assumption to guarantee robustness and the confidentiality of updates. It is, therefore, currently impossible to ensure Byzantine robustness and confidentiality of updates without assuming a semi-honest majority. To tackle this problem, we propose a novel Byzantine-robust and privacy-preserving FL system, called MUDGUARD, to capture malicious minority and majority for server and client sides, respectively. Our experimental results demonstrate that the accuracy of MUDGUARD is practically close to the FL baseline using FedAvg without attacks (≈0.8% gap on average). Meanwhile, the attack success rate is around 0%-5% even under an adaptive attack tailored to MUDGUARD. We further optimize our design by using binary secret sharing and polynomial transformation, leading to communication overhead and runtime decreases of 67%-89.17% and 66.05%-68.75%, respectively.