INRIA a CCSD electronic archive server
Not a member yet
122212 research outputs found
Sort by
Combined Deep KKL and NMPC for Modeling Muscle Co-contraction Under Measurement Delays
No full textThis paper addresses the modeling of robust control through muscle co-contraction, a fundamental physiological strategy that modulates mechanical impedance and stabilizes limb dynamics. Here, the neuromusculoskeletal system is modeled as nonlinear dynamics actuated by antagonist muscles, subject to uncertainties and significant output measurement delays. To account for muscle co-contraction for robust control in simulation, we propose a closed-loop control framework based on Nonlinear Model Predictive Control (NMPC) and test it for a forearm stabilization task (inverted pendulum). While NMPC is well-suited for handling system constraints and mimicking physiological objectives, its real-time implementation requires the estimation of the instantaneous initial state, which is unavailable due to the delays. To address this, we introduce a delay-compensating Kazantzis-Kravaris-Luenberger (KKL) observer. This observer embeds the nonlinear dynamics into a stable linear latent space, enabling explicit delay compensation through a chain of predictors. To overcome the analytical complexity of the required transformation maps, we employ a Deep Learning approach to approximate the observer dynamics. Finally, the performance and robustness of the Deep KKL-NMPC scheme are validated through extensive simulations of posture stabilization and dynamic tracking tasks under stochastic environmental disturbances
k-scale: k-Anonymizing Millions of Trajectories
No full textInternational audienceTrajectory datasets collected by network operators and service providers offer detailed information about individual mobility and have wide application in business and research. However, managing such data raises privacy risks, as the unique movement patterns of individuals pose significant re-identification risks and make common countermeasures like pseudonymization ineffective. The privacy-preserving data publishing (PPDP) of trajectory datasets that maintains post-anonymization accuracy and truthfulness is an open problem -especially for large datasets with millions of records like those gathered by major actors in the telco ecosystem. We close this gap with k-scale, a framework that implements k-anonymity in massive mobile user trajectory datasets, removing uniqueness while safeguarding accuracy at the record level. Not only k-scale is the first model capable of scaling k-anonymization to a dataset of one million trajectories, but it does so while also outperforming state-of-the-art methods for trajectory data publishing in terms of preserved data quality, which we prove in real-world massive datasets and applications
Dynamic Agent Generation for Self-Adaptive Root Cause Analysis
No full textInternational audienceModern software systems increasingly rely on microservice architectures, which enhance modularity and resilience but produce vast amounts of heterogeneous observability data—logs, metrics, and traces—to ensure reliable operation and early failure detection. Performing Root Cause Analysis (RCA) on such data is challenging due to its scale, heterogeneity, and evolving structure, which hinder effective correlation and reasoning across modalities. Although recent studies have explored statistical techniques, graph-based models, and Large Language Models (LLMs)-based agents for RCA, most remain static and task-specific, lacking the adaptability and coordination required to handle evolving diagnostic contexts. This paper introduces a self-adaptive agent generation framework that leverages LLMs to dynamically compose and orchestrate adapted diagnostic agents at runtime according to each anomaly’s characteristics. Two main agents drive this process: a Parser Agent that interprets natural-language queries and builds structured task specifications, and an Executor Agent that adapts and coordinates adapted agents analyzing multimodal observability data through a shared memory space. Experiments on Nezha (fault-injected) and OpenRCA (real-world) datasets show up to 12% and 22% gains in diagnostic accuracy, confirming the framework’s effectiveness in adaptive reasoning, coordination, and interpretable root cause identification
DUALF-D: Disentangled dual-hyperprior approach for light field image compression
No full textInternational audienceLight field (LF) imaging captures spatial and angular information, offering a 4D scene representation enabling enhanced visual understanding. However, high dimensionality and redundancy across spatial and angular domains present major challenges for compression, particularly where storage, transmission bandwidth, or processing latency are constrained. We present a novel Variational Autoencoder (VAE)-based framework that explicitly disentangles spatial and angular features using two parallel latent branches. Each branch is coupled with an independent hyperprior model, allowing more precise distribution estimation for entropy coding and finer rate-distortion control. This dual-hyperprior structure enables the network to adaptively compress spatial and angular information based on their unique statistical characteristics, improving coding efficiency. To further enhance latent feature specialization and promote disentanglement, we introduce a mutual information-based regularization term that minimizes redundancy between the two branches while preserving feature diversity. Unlike prior methods relying on covariance-based penalties prone to collapse, our information-theoretic regularizer provides more stable and interpretable latent separation. Experimental results on publicly available LF datasets demonstrate our method achieves strong compression performance, yielding an average BD-PSNR gain of 2.91 dB over HEVC and high compression ratios (e.g., 200:1). Additionally, our design enables fast inference, with a total end-to-end time over 19x faster than the JPEG Pleno standard, making it well-suited for real-time and bandwidth-sensitive applications. By jointly leveraging disentangled representation learning, dual-hyperprior modeling, and information-theoretic regularization, our approach offers a scalable, effective solution for practical light field image compression.</div
ORKM: An R package for online multi-view data clustering
No full textInternational audienceWe propose a package called ORKM, which implements the ORKMC (Online Regularized K-Means Clustering) method for handling online multi-view or single-view data, which named ORKMeans function in the package incorporates a regularization term to address multi-view clustering problems with online updates. ORKM computes classification results, cluster center matrices, and view-specific weights for multi-view datasets. It also supports branching multi/single-view data by converting the online RKMC algorithm into an offline version, referred to as RKMC (Regularized K-Means Clustering) realized by function RKMeans. We demonstrate the package’s functionality through simulations and real-world data analyses, comparing it with other methods and related R packages. Overall, ORKM exhibits stable performance and effective clustering result
Efficient Threshold ML-DSA
No full textInternational audienceThreshold signature schemes allow a group of users to jointly generate a digital signature, providing resilience against faults and enhancing decentralization. With the advent of post-quantum cryptography, lattice-based threshold signatures have gained attention as viable PQ-threshold solutions. Nevertheless, existing constructions are limited in terms of their scalability, robustness. Worse, none is compatible with standardized schemes, particularly with the NIST-selected and standardized Module-Lattice-based Digital Signature Algorithm (ML-DSA) algorithm.In this work, we present the first threshold signature scheme that is fully compatible with ML-DSA, supporting secure and efficient signing for a small number of parties, with an average communication per party upper bounded by 1 MB up to 6 parties. Our construction leverages advanced short secret sharing techniques and integrates optimized rejection sampling to achieve a favorable balance between communication efficiency and correctness in distributed environments. We implement our construction in Go and evaluate its performance across local, LAN, and WAN network settings. Our benchmarks demonstrate that our threshold ML-DSA scheme is not only practically deployable but also well-suited for real-world applications, including multi-device cryptocurrency wallets, threshold-based TLS authentication, and for Tor's directory authorities
Towards Pen-and-Paper-Style Equational Reasoning in Interactive Theorem Provers by Equality Saturation
No full textInternational audienceEquations are ubiquitous in mathematical reasoning. Often, however, they only hold under certain conditions. As these conditions are usually clear from context, mathematicians regularly omit them when performing equational reasoning on paper. In contrast, interactive theorem provers pedantically insist on every detail to be convinced that a theorem holds, hindering equational reasoning at the more abstract level of pen-andpaper mathematics. In this paper, we address this issue by raising the level of equational reasoning to enable pen-and-paper style in interactive theorem provers. We achieve this by interpreting theorems as conditional rewrite rules, and use equality saturation to automatically derive equational proofs. Conditions that cannot be automatically proven may be surfaced as proof obligations. Concretely, we present how to interpret theorems as conditional rewrite rules for a significant class of theorems. Handling these theorems goes beyond simple syntactic rewriting, and deals with aspects like propositional conditions and type classes. We evaluate our approach by implementing it as a tactic in Lean, using the egg library for equality saturation with e-graphs. We show four use cases demonstrating the efficacy of this higher level of abstraction for equational reasoning
DroidHunter : une détection robuste basée sur la vision contre les logiciels malveillants Android obfusqués
No full textInternational audienceDue to their large popularity, Android smartphones are often targeted by malware attacks. Several strategies exist to detect malwarecode. However, we show that they are insufficient when dealingwith obfuscation techniques. DroidHunter is our novel methodfor detecting Android malwares. DroidHunter leverages opcodesand their parameters, transforming those into RGB images andspecific encoding techniques. The generated images are then usedto train two different classification models based on support vector machines, convolutional neural networks, and a vision-basedtransformer. We evaluate DroidHunter on several datasets withup to 476,937 APKs from multiple sources. With detection ratesfrom 98.65% to 99.94%, DroidHunter overcomes nine state-of-the-art malware detection techniques, including Drebin, MaMadroid,DexRay. Moreover, DroidHunter demonstrates strong resilienceagainst hidden malware with detection rates up to 98.98%, andshows robustness on newly emerging threats, achieving an AUT of0.89 on recent malware samples. We release our code to the researchcommunity, with instructions to reproduce our evaluation availableat: https://zenodo.org/doi/10.5281/zenodo.10977166.En raison de leur grande popularité, les smartphones Android sont souvent la cible d'attaques de logiciels malveillants. Il existe plusieurs stratégies pour détecter les codes malveillants. Cependant, nous montrons qu'elles sont insuffisantes lorsqu'il s'agit de techniques d'obfuscation. DroidHunter est notre nouvelle méthode de détection des logiciels malveillants Android. DroidHunter exploite les codes opérationnels et leurs paramètres, en les transformant en images RVB et en techniques d'encodage spécifiques. Les images générées sont ensuite utilisées pour entraîner deux modèles de classification différents basés sur des machines à vecteurs de support, des réseaux neuronaux convolutifs et un transformateur basé sur la vision. Nous évaluons DroidHunter sur plusieurs ensembles de données comprenant jusqu'à 476 937 APK provenant de multiples sources. Avec des taux de détection compris entre 98,65 % et 99,94 %, DroidHunter surpasse neuf techniques de détection de logiciels malveillants de pointe, notamment Drebin, MaMadroid etACM Asia Conference on Computer and Communications Security (ASIA CCS '26)DexRay. De plus, DroidHunter fait preuve d'une forte résilienceface aux logiciels malveillants cachés, avec des taux de détection pouvant atteindre 98,98 %, et se montre robuste face aux nouvelles menaces émergentes, avec un AUT de 0,89 sur des échantillons de logiciels malveillants récents. Nous publions notre code à la communauté des chercheurs, avec des instructions pour reproduire notre évaluation disponibles à l'adresse suivante : https://zenodo.org/doi/10.5281/zenodo.10977166
Bigraded Castelnuovo-Mumford regularity and Groebner bases
No full textInternational audienceWe study the relation between the bigraded Castelnuovo-Mumford regularity of abihomogeneous ideal in the coordinate ring of the product of two projective spaces and the bidegrees of a Groebner basis of with respect to the degree reverse lexicographical monomial order in generic coordinates. For the single-graded case, Bayer and Stillman unraveled all aspects of this relationship forty years ago and these results led to complexity estimates for computations with Groebner bases. We build on this work to introduce a bounding region of the bidegrees of minimal generators of bihomogeneous Groebner bases for . We also use this region to certify the presence of some minimalgenerators close to its boundary. Finally, we show that, up to a certain shift, this region is related to the bigraded Castelnuovo-Mumford regularity of
Ergodicity of some stochastic Fokker-Planck equations with additive common noise
No full textInternational audienceIn this paper we consider stochastic Fokker-Planck Partial Differential Equations (PDEs), obtained as the mean-field limit of weakly interacting particle systems subjected to both independent (or idiosyncratic) and common Brownian noises. We provide sufficient conditions under which the deterministic counterpart of the Fokker-Planck equation, which corresponds to particle systems that are just subjected to independent noises, has several invariant measures, but for which the stochastic version admits a unique invariant measure under the presence of the additive common noise. The very difficulty comes from the fact that the common noise is just of finite dimension while the state variable, which should be seen as the conditional marginal law of the system given the common noise, lives in a space of infinite dimension. In this context, our result holds true if, in addition to standard confining properties, the mean field interaction term forces the system to be attracted by its conditional mean given the common noise and the intensity of the idiosyncratic noise is small