arXiv subjects

Sonia Ben Mokhtar

Publications and source records attributed to Sonia Ben Mokhtar.

At least 19 recordsLinked to original sources

arXiv2026

Giskard : Byzantine Robust and Confidential Aggregation for Large-Scale Decentralized Learning

Dealing simultaneously with confidentiality and Byzantine behaviors in decentralized learning is a challenging problem. Indeed, in decentralized learning, clients train a machine learning model while keeping their data locally and share their model parameters or gradients with a set of neighbors. While enforcing confidentiality calls for hiding the exchanged model parameters/gradients (e.g., by using cryptographic techniques), dealing with Byzantine contributions often requires inspecting the latter. Hence, most research works address these objectives separately. A recent line of work proposes to employ secure multi-party computation (MPC) to implement robust aggregators against model poisoning, thereby enforcing both confidentiality and Byzantine resilience. However, these solutions scale badly: they either require all-to-all communication between participants or delegate the entire computation to a small subset, whose computational and communication load grows proportionally with the size of the network. In this paper, we present Giskard, a protocol for confidential and Byzantine-robust decentralized aggregation. Giskard organizes $n$ parties into a tree of committees of size $O(\log n)$ and evaluates a coordinate-wise approximate median via a committee-adapted distributed binary search over the value domain, using BGW-style MPC within each committee. We assess Giskard both theoretically by proving its security and confidentiality properties and experimentally through extensive experiments involving up to one million participants. Compared to its closest competitors, Giskard reduces per-party communication complexity asymptotically while exhibiting comparable model utility under up to $n/4$ Byzantine parties.

cs.CR↗

arXiv2026

GRANITE : a Byzantine-Resilient Dynamic Gossip Learning Framework

Gossip Learning (GL) is a decentralized learning paradigm where users iteratively exchange and aggregate models with a small set of neighboring peers. Recent approaches rely on dynamic communication graphs built using Random Peer Sampling (RPS) protocols which have been proven to accelerate convergence. However, we show that these approaches are vulnerable to a dual attack: Byzantine nodes can poison models and manipulate peer sampling to amplify their influence. We address this combination of threats with GRANITE, a framework for robust learning over sparse, dynamic graphs in the presence of Byzantine nodes. GRANITE accumulates knowledge about encountered node identifiers over time and dynamically adjusts local aggregation thresholds based on estimated Byzantine density in the neighbourhood of each node. We demonstrate that under GRANITE, the Byzantine presence in local neighborhoods exhibits an exponential decay. We further derive the robustness conditions of the graphs generated by GRANITE. Empirically, our results indicate that GRANITE converges within 5% of non-Byzantine accuracy under 30% Byzantines nodes, offers faster convergence and operates on graphs with up to 9x lower communication cost.

cs.LG↗

arXiv2026

On the Normalization of Confusion Matrices: Methods and Geometric Interpretations

The confusion matrix is a standard tool for evaluating classifiers by providing insights into class-level errors. In heterogeneous settings, its values are shaped by two main factors: class similarity -- how easily the model confuses two classes -- and distribution bias, arising from skewed distributions in the training and test sets. However, confusion matrix values reflect a mix of both factors, making it difficult to disentangle their individual contributions. To address this, we introduce bistochastic normalization using Iterative Proportional Fitting, a generalization of row and column normalization. Unlike standard normalizations, this method recovers the underlying structure of class similarity. By disentangling error sources, it enables more accurate diagnosis of model behavior and supports more targeted improvements. We also show a correspondence between confusion matrix normalizations and the model's internal class representations. Both standard and bistochastic normalizations can be interpreted geometrically in this space, offering a deeper understanding of what normalization reveals about a classifier.

cs.LG↗

arXiv2026

DOME: Improving Signal-to-Noise in Stochastic Gradient Descent via Sharp-Direction Subspace Filtering

Stochastic gradients for deep neural networks exhibit strong correlations along the optimization trajectory, and are often aligned with a small set of Hessian eigenvectors associated with outlier eigenvalues. Recent work shows that projecting gradients away from this Hessian outlier subspace has little impact on optimization, despite capturing a large fraction of gradient variability. Since computing the Hessian is intractable in practice, we introduce a principled first-order characterization of the nuisance subspace based on the covariance of stochastic gradients, and propose an efficient method to estimate it online. We show that removing this subspace also has little impact on optimization, and yields practical benefits for applications sensitive to gradient signal-to-noise ratio such as gradient compression.

cs.LG↗

arXiv2025

TriHaRd: Higher Resilience for TEE Trusted Time

Accurately measuring time passing is critical for many applications. However, in Trusted Execution Environments (TEEs) such as Intel SGX, the time source is outside the Trusted Computing Base: a malicious host can manipulate the TEE's notion of time, jumping in time or affecting perceived time speed. Previous work (Triad) proposes protocols for TEEs to maintain a trustworthy time source by building a cluster of TEEs that collaborate with each other and with a remote Time Authority to maintain a continuous notion of passing time. However, such approaches still allow an attacker to control the operating system and arbitrarily manipulate their own TEE's perceived clock speed. An attacker can even propagate faster passage of time to honest machines participating in Triad's trusted time protocol, causing them to skip to timestamps arbitrarily far in the future. We propose TriHaRd, a TEE trusted time protocol achieving high resilience against clock speed and offset manipulations, notably through Byzantine-resilient clock updates and consistency checks. We empirically show that TriHaRd mitigates known attacks against Triad.

cs.CR↗

arXiv2025

Exposing the Vulnerability of Decentralized Learning to Membership Inference Attacks Through the Lens of Graph Mixing

The primary promise of decentralized learning is to allow users to engage in the training of machine learning models in a collaborative manner while keeping their data on their premises and without relying on any central entity. However, this paradigm necessitates the exchange of model parameters or gradients between peers. Such exchanges can be exploited to infer sensitive information about training data, which is achieved through privacy attacks (e.g., Membership Inference Attacks -- MIA). In order to devise effective defense mechanisms, it is important to understand the factors that increase/reduce the vulnerability of a given decentralized learning architecture to MIA. In this study, we extensively explore the vulnerability to MIA of various decentralized learning architectures by varying the graph structure (e.g., number of neighbors), the graph dynamics, and the aggregation strategy, across diverse datasets and data distributions. Our key finding, which to the best of our knowledge we are the first to report, is that the vulnerability to MIA is heavily correlated to (i) the local model mixing strategy performed by each node upon reception of models from neighboring nodes and (ii) the global mixing properties of the communication graph. We illustrate these results experimentally using four datasets and by theoretically analyzing the mixing properties of various decentralized architectures. We also empirically show that enhancing mixing properties is highly beneficial when combined with other privacy-preserving techniques such as Differential Privacy. Our paper draws a set of lessons learned for devising decentralized learning systems that reduce by design the vulnerability to MIA.

cs.LG↗

arXiv2025

A Weighted Loss Approach to Robust Federated Learning under Data Heterogeneity

Federated learning (FL) is a machine learning paradigm that enables multiple data holders to collaboratively train a machine learning model without sharing their training data with external parties. In this paradigm, workers locally update a model and share with a central server their updated gradients (or model parameters). While FL seems appealing from a privacy perspective, it opens a number of threats from a security perspective as (Byzantine) participants can contribute poisonous gradients (or model parameters) harming model convergence. Byzantine-resilient FL addresses this issue by ensuring that the training proceeds as if Byzantine participants were absent. Towards this purpose, common strategies ignore outlier gradients during model aggregation, assuming that Byzantine gradients deviate more from honest gradients than honest gradients do from each other. However, in heterogeneous settings, honest gradients may differ significantly, making it difficult to distinguish honest outliers from Byzantine ones. In this paper, we introduce the Worker Label Alignement Loss (WoLA), a weighted loss that aligns honest worker gradients despite data heterogeneity, which facilitates the identification of Byzantines' gradients. This approach significantly outperforms state-of-the-art methods in heterogeneous settings. In this paper, we provide both theoretical insights and empirical evidence of its effectiveness.

cs.LG↗

arXiv2025

COoL-TEE: Client-TEE Collaboration for Resilient Distributed Search

Current marketplaces rely on search mechanisms with distributed systems but centralized governance, making them vulnerable to attacks, failures, censorship and biases. While search mechanisms with more decentralized governance (e.g., DeSearch) have been recently proposed, these are still exposed to information head-start attacks (IHS) despite the use of Trusted Execution Environments (TEEs). These attacks allow malicious users to gain a head-start over other users for the discovery of new assets in the market, which give them an unfair advantage in asset acquisition. We propose COoL-TEE, a TEE-based provider selection mechanism for distributed search, running in single- or multi-datacenter environments, that is resilient to information head-start attacks. COoL-TEE relies on a Client-TEE collaboration, which enables clients to distinguish between slow providers and malicious ones. Performance evaluations in single- and multi-datacenter environments show that, using COoL-TEE, malicious users respectively gain only up to 2% and 7% of assets more than without IHS, while they can claim 20% or more on top of their fair share in the same conditions with DeSearch.

cs.DC↗

arXiv2025

An Open-source Implementation and Security Analysis of Triad's TEE Trusted Time Protocol

The logic of many protocols relies on time measurements. However, in Trusted Execution Environments (TEEs) like Intel SGX, the time source is outside the Trusted Computing Base: a malicious system hosting the TEE can manipulate that TEE's notion of time, e.g., jumping in time or affecting the perceived time speed. Previous work like Triad propose protocols for TEEs to maintain a trustworthy time source. However, in this paper, based on a public implementation of Triad that we contribute, we empirically showcase vulnerabilities to this protocol. For example, an attacker controlling the operating system, and consequently the scheduling algorithm, may arbitrarily manipulate their local TEE's clock speed. What is worse, in case of faster malicious clock speeds, an attacker on a single compromised machine may propagate the attack to honest machines participating in Triad's Trusted Time protocol, causing them to skip to timestamps arbitrarily far in the future. Then, infected honest machines propagate time-skips themselves to other honest machines interacting with them. We discuss protocol changes to Triad for higher resilience against such attacks.

cs.CR↗

arXiv2025

Differentially private and decentralized randomized power method

The randomized power method has gained significant interest due to its simplicity and efficient handling of large-scale spectral analysis and recommendation tasks. However, its application to large datasets containing personal information (e.g., web interactions, search history, personal tastes) raises critical privacy problems. This paper addresses these issues by proposing enhanced privacy-preserving variants of the method. First, we propose a variant that reduces the amount of the noise required in current techniques to achieve Differential Privacy (DP). More precisely, we refine the privacy analysis so that the Gaussian noise variance no longer grows linearly with the target rank, achieving the same DP guarantees with strictly less noise. Second, we adapt our method to a decentralized framework in which data is distributed among multiple users. The decentralized protocol strengthens privacy guarantees with no accuracy penalty and a low computational and communication overhead. Our results include the provision of tighter convergence bounds for both the centralized and decentralized versions, and an empirical comparison with previous work using real recommendation datasets.

cs.LG↗

arXiv2025

Dropout-Robust Mechanisms for Differentially Private and Fully Decentralized Mean Estimation

Achieving differentially private computations in decentralized settings poses significant challenges, particularly regarding accuracy, communication cost, and robustness against information leakage. While cryptographic solutions offer promise, they often suffer from high communication overhead or require centralization in the presence of network failures. Conversely, existing fully decentralized approaches typically rely on relaxed adversarial models or pairwise noise cancellation, the latter suffering from substantial accuracy degradation if parties unexpectedly disconnect. In this work, we propose IncA, a new protocol for fully decentralized mean estimation, a widely used primitive in data-intensive processing. Our protocol, which enforces differential privacy, requires no central orchestration and employs low-variance correlated noise, achieved by incrementally injecting sensitive information into the computation. First, we theoretically demonstrate that, when no parties permanently disconnect, our protocol achieves accuracy comparable to that of a centralized setting-already an improvement over most existing decentralized differentially private techniques. Second, we empirically show that our use of low-variance correlated noise significantly mitigates the accuracy loss experienced by existing techniques in the presence of dropouts.

cs.CR↗

arXiv2025

Inferring Communities of Interest in Collaborative Learning-based Recommender Systems

Collaborative-learning-based recommender systems, such as those employing Federated Learning (FL) and Gossip Learning (GL), allow users to train models while keeping their history of liked items on their devices. While these methods were seen as promising for enhancing privacy, recent research has shown that collaborative learning can be vulnerable to various privacy attacks. In this paper, we propose a novel attack called Community Inference Attack (CIA), which enables an adversary to identify community members based on a set of target items. What sets CIA apart is its efficiency: it operates at low computational cost by eliminating the need for training surrogate models. Instead, it uses a comparison-based approach, inferring sensitive information by comparing users' models rather than targeting any specific individual model. To evaluate the effectiveness of CIA, we conduct experiments on three real-world recommendation datasets using two recommendation models under both Federated and Gossip-like settings. The results demonstrate that CIA can be up to 10 times more accurate than random guessing. Additionally, we evaluate two mitigation strategies: Differentially Private Stochastic Gradient Descent (DP-SGD) and a Share less policy, which involves sharing fewer, less sensitive model parameters. Our findings suggest that the Share less strategy offers a better privacy-utility trade-off, especially in GL.

cs.IR↗

arXiv2025

Reliability is Blind: Collective Incentives for Decentralized Computing Marketplaces without Individual Behavior Information

In decentralized cloud computing marketplaces, ensuring fair and efficient interactions among asset providers and end-users is crucial. A key concern is meeting agreed-upon service-level objectives like the service's reliability. In this decentralized context, traditional mechanisms often fail to address the complexity of task failures, due to limited available and trustworthy insights into these independent actors' individual behavior. This paper proposes a collective incentive mechanism that blindly punishes all involved parties when a task fails. Based on ruin theory, we show that Collective Incentives improve behavior in the marketplace by creating a disincentive for faults and misbehavior even when the parties at fault are unknown, in turn leading to a more robust marketplace. Simulations for small and large pools of marketplace assets show that Collective Incentives enable to meet or exceed a reliability target, i.e., the success-rate of tasks run using marketplace assets, by eventually discarding failure-prone assets while preserving reliable ones.

cs.DC↗

arXiv2025

Secure Federated Graph-Filtering for Recommender Systems

Recommender systems often rely on graph-based filters, such as normalized item-item adjacency matrices and low-pass filters. While effective, the centralized computation of these components raises concerns about privacy, security, and the ethical use of user data. This work proposes two decentralized frameworks for securely computing these critical graph components without centralizing sensitive information. The first approach leverages lightweight Multi-Party Computation and distributed singular vector computations to privately compute key graph filters. The second extends this framework by incorporating low-rank approximations, enabling a trade-off between communication efficiency and predictive performance. Empirical evaluations on benchmark datasets demonstrate that the proposed methods achieve comparable accuracy to centralized state-of-the-art systems while ensuring data confidentiality and maintaining low communication costs. Our results highlight the potential for privacy-preserving decentralized architectures to bridge the gap between utility and user data protection in modern recommender systems.

cs.IR↗

arXiv2025

TEE-based Key-Value Stores: a Survey

Key-Value Stores (KVSs) are No-SQL databases that store data as key-value pairs and have gained popularity due to their simplicity, scalability, and fast retrieval capabilities. However, storing sensitive data in KVSs requires strong security properties to prevent data leakage and unauthorized tampering. While software (SW)-based encryption techniques are commonly used to maintain data confidentiality and integrity, they suffer from several drawbacks. They strongly assume trust in the hosting system stack and do not secure data during processing unless using performance-heavy techniques (e.g., homomorphic encryption). Alternatively, Trusted Execution Environments (TEEs) provide a solution that enforces the confidentiality and integrity of code and data at the CPU level, allowing users to build trusted applications in an untrusted environment. They also secure data in use by providing an encapsulated processing environment called enclave. Nevertheless, TEEs come with their own set of drawbacks, including performance issues due to memory size limitations and CPU context switching. This paper examines the state of the art in TEE-based confidential KVSs and highlights common design strategies used in KVSs to leverage TEE security features while overcoming their inherent limitations. This work aims to provide a comprehensive understanding of the use of TEEs in KVSs and to identify research directions for future work.

cs.CR↗

arXiv2024

Survey of Federated Learning Models for Spatial-Temporal Mobility Applications

Federated learning involves training statistical models over edge devices such as mobile phones such that the training data is kept local. Federated Learning (FL) can serve as an ideal candidate for training spatial temporal models that rely on heterogeneous and potentially massive numbers of participants while preserving the privacy of highly sensitive location data. However, there are unique challenges involved with transitioning existing spatial temporal models to decentralized learning. In this survey paper, we review the existing literature that has proposed FL-based models for predicting human mobility, traffic prediction, community detection, location-based recommendation systems, and other spatial-temporal tasks. We describe the metrics and datasets these works have been using and create a baseline of these approaches in comparison to the centralized settings. Finally, we discuss the challenges of applying spatial-temporal models in a decentralized setting and by highlighting the gaps in the literature we provide a road map and opportunities for the research community.

cs.LG↗

arXiv2022

Shielding Federated Learning Systems against Inference Attacks with ARM TrustZone

Federated Learning (FL) opens new perspectives for training machine learning models while keeping personal data on the users premises. Specifically, in FL, models are trained on the users devices and only model updates (i.e., gradients) are sent to a central server for aggregation purposes. However, the long list of inference attacks that leak private data from gradients, published in the recent years, have emphasized the need of devising effective protection mechanisms to incentivize the adoption of FL at scale. While there exist solutions to mitigate these attacks on the server side, little has been done to protect users from attacks performed on the client side. In this context, the use of Trusted Execution Environments (TEEs) on the client side are among the most proposing solutions. However, existing frameworks (e.g., DarkneTZ) require statically putting a large portion of the machine learning model into the TEE to effectively protect against complex attacks or a combination of attacks. We present GradSec, a solution that allows protecting in a TEE only sensitive layers of a machine learning model, either statically or dynamically, hence reducing both the TCB size and the overall training time by up to 30% and 56%, respectively compared to state-of-the-art competitors.

cs.CR↗

arXiv2022

PEPPER: Empowering User-Centric Recommender Systems over Gossip Learning

Recommender systems are proving to be an invaluable tool for extracting user-relevant content helping users in their daily activities (e.g., finding relevant places to visit, content to consume, items to purchase). However, to be effective, these systems need to collect and analyze large volumes of personal data (e.g., location check-ins, movie ratings, click rates .. etc.), which exposes users to numerous privacy threats. In this context, recommender systems based on Federated Learning (FL) appear to be a promising solution for enforcing privacy as they compute accurate recommendations while keeping personal data on the users' devices. However, FL, and therefore FL-based recommender systems, rely on a central server that can experience scalability issues besides being vulnerable to attacks. To remedy this, we propose PEPPER, a decentralized recommender system based on gossip learning principles. In PEPPER, users gossip model updates and aggregate them asynchronously. At the heart of PEPPER reside two key components: a personalized peer-sampling protocol that keeps in the neighborhood of each node, a proportion of nodes that have similar interests to the former and a simple yet effective model aggregation function that builds a model that is better suited to each user. Through experiments on three real datasets implementing two use cases: a location check-in recommendation and a movie recommendation, we demonstrate that our solution converges up to 42% faster than with other decentralized solutions providing up to 9% improvement on average performance metric such as hit ratio and up to 21% improvement on long tail performance compared to decentralized competitors.

cs.IR↗