Portail HAL Ensta
Not a member yet
11080 research outputs found
Sort by
Interpretable Rule Learning for Reactive Activity Recognition in Event-Driven RL Environments
No full textInternational audienceMany real-world environments, from smart homes to industrial systems, produce asynchronous event streams driven by latent activities with complex temporal structures. Recognizing these patterns requires reasoning over temporal dependencies that reactive policies alone do not capture. We propose a hybrid reinforcement learning framework that combines symbolic program synthesis with reactive policy optimization for interpretable activity recognition. This hybrid approach enables the agent to disambiguate overlapping activities, generalize across history patterns, and maintain interpretable decision logic. Our method discovers temporal rules as logical abstractions over event histories, using a compositional grammar based on Allen's interval algebra. Monte Carlo Tree Search (MCTS) explores the rule space, refining candidates to maximize cumulative reward. The resulting rules define symbolic contexts that augment the observable state and support decision-making in a near-Markovian surrogate process. Evaluations on a synthetic benchmark with concurrent, asynchronous activities show strong task performance and symbolic fidelity compared to neural and evolutionary baselines
The emergence of 4th Industrial Revolution technologies in defense innovation in the Indo-Pacific region
No full textInternational audienc
Attraction of the core and the cohesion flow
No full textInternational audienceWe adopt a continuous-time dynamical system approach to study the evolution of the state of a game driven by the willingness to reduce the total dissatisfaction of the coalitions about their payment. Inspired by the work of Grabisch and Sudhölter about core stability, we define a vector field on the set of preimputations from which is defined, for any preimputation, a cohesion curve describing the evolution of the state. We prove that for each preimputation, there exists a unique cohesion curve. Subsequently, we show that, for the cohesion flow of a balanced game, the core is the unique minimal attractor of the flow, the realm of which is the whole preimputation set. These results improve our understanding of the ubiquity of the core in the study of cooperative games with transferable utility
Asymptotic approaches for dealing with distorted crack geometries
No full textA growing, albeit not predominant, way to contribute to the climate transition in the field of fracturemechanics is to refine predictions from damage tolerance approaches used to assess the durability andreliability of sensitive components such as those found in railway, aeronautics, aerospace or nuclearindustries. If it can be certified that the reduced safety margins remain acceptable, the intervals betweenmaintenance operations could be extended, and the replacement of defective parts delayed, therebyreducing the environmental footprint.The aim of Damage Tolerance Approaches is to ensure that the existence of unavoidable defects does notcompromise safety. This involves considering the most unfavorable case of brittle fracture and determiningthe propagation of a preexisting crack under cyclic loading until the Griffith energy fracture thresholdis reached. Currently, this is done using simplified, smoothed-out crack geometries, relying either ontabulated values or Finite Element Methods. Since meshing of the entire structure is required, the lattercurrently struggles to accurately account for the small-scale tortuosity of the crack geometry.This paper aims to show that asymptotic approaches are efficient alternatives to address this challenge.Various aspects of these approaches, along with selected applications, will be reviewed. In addition tosupporting the reduction of safety margins, these approaches also help to ensure that small geometricalperturbations do not lead to unexpected catastrophic failure. Asymptotic methods aim to provide analyticalformulas for the variation of stress intensity factors caused by small-scale crack shape perturbations.Several applications are presented including crack shape evolution, the influence of heterogeneities, andpropagation in general mixed-mode I+II+III condition
Minutes from the 9 th edition of the Performance and Scalability of Storage Systems workshop (Per3S), 23rd May 2025, "Maison des Mines et des Ponts", Paris
No full textInternational audiencePer3S is a workshop aiming to bring together the scientific and technological storage community to discuss and address issues and challenges associated to performance and data operations at scale. These topics cover HPC storage as well as Cloud-oriented architectures, both sharing the need for extreme scale
Nonlinear manifold approximation using compositional polynomial networks
No full text26 pages, 9 figuresWe consider the problem of approximating a subset of a Hilbert space by a low-dimensional manifold , using samples from . We propose a nonlinear approximation method where is defined as the range of a smooth nonlinear decoder defined on with values in a possibly high-dimensional linear space , and a linear encoder which associates to an element from its coefficients on a basis of a -dimensional subspace , where and are optimal or near to optimal linear spaces, depending on the selected error measure. The linearity of the encoder allows to easily obtain the parameters associated with a given element in . The proposed decoder is a polynomial map from to which is obtained by a tree-structured composition of polynomial maps, estimated sequentially from samples in . Rigorous error and stability analyses are provided, as well as an adaptive strategy for constructing a decoder that guarantees an approximation of the set with controlled mean-squared or wort-case errors, and a controlled stability (Lipschitz continuity) of the encoder and decoder pair. We demonstrate the performance of our method through numerical experiments
Enhancing Keystone Security Against Cache Timing Attacks: A Modular Approach
No full textInternational audienceConfidential computing includes various methods to enhance data security, notably by processing sensitive information within Trusted Execution Environments (TEEs). However, TEEs remain vulnerable to Side-Channel Attacks (SCAs), such as cache timing attacks, which exploit timing variations to extract confidential data. Existing TEE designs do not provide sufficient protection against these threats, highlighting the need for stronger security measures. This study focuses on integrating countermeasures specifically targeting timing and cache vulnerabilities within a TEE. The implementation will leverage the RISC-V architecture to explore its potential in mitigating SCA within TEE
Visco-elasto-plastic characterization and modeling of a wet polyamide laid-strand sub-rope for floating offshore wind turbine moorings
No full textInternational audiencePolyamide 6 fiber ropes are of interest for floating offshore wind turbine mooring lines but could exhibit complex mechanical behavior during loading at sea, such as creep, relaxation, variable dynamic stiffness or visco-plasticity. There is a need for a model that could be introduced into finite element analyses to predict this complex response; it should also describe the effect of the loading history. This paper presents a visco-elasto-plastic behavior model based on four dashpot-ratchet-spring elements that allow a precise description of polyamide 6 rope behavior. An identification method, using a multi-relaxation test, is described. It has been implemented in a finite element analysis software and validations are made by comparing the model results to the experimental data. The present work is the result of an extensive effort initiated by the collaborative research project POLYAMOOR and continued by the MONAMOOR project, both led by France Energies Marines
Les inégalités d’intégration au sein de l’espace économique méditerranéen au tournant du xxie siècle
No full textInternational audienceWe propose a new indicator to measure inequalities in the integration of countries within a trade area. This indicator combines a global approach with another revealing the specific contribution of each country to its value. We apply it to the Mediterranean economic area (MEA) over the period 1980-2004, a period that preceded an intense institutional dynamic aimed at further integrating the North and South of the Mediterranean. Prior to these institutional transformations, the EEM remained a relatively unintegrated zone overall, and this low level of integration was maintained by the main European countries.Nous proposons un nouvel indicateur mesurant les inégalités d’intégration des pays au sein d’un espace d’échanges commerciaux. Cet indicateur associe une approche globale à une autre révélant la contribution spécifique de chaque pays à sa valeur. Nous l’appliquons à l’espace économique méditerranéen (EEM) sur la période 1980-2004, période précédant une intense dynamique institutionnelle visant à intégrer davantage le Nord et le Sud de la Méditerranée. Avant ces transformations institutionnelles, l’EEM reste une zone globalement peu intégrée, et cette faible intégration est entretenue par les principaux pays d’Europe qui en sont membres
Modeling and control of an hoverboard
No full textIn this paper, we explain how to control an hoverboard in a two dimensional world. This means that the hoverboard can only move forward and backward on a line. An extension to a three dimensional hoverboard with a heading control is not treated here. Our planar hoverboard has only one actuator (a motor in the axle) and one inertial unit (for the acceleration and the gyrometer). Since the hoverboard has no idea of the position, the weight, and the size of the human rider, a model free approach is considered