HAL Portal IOGS (nstitut d'Optique Graduate School)
Not a member yet
12589 research outputs found
Sort by
Radiation-induced effects on commercial 3D printing materials exposed to high x-ray doses
No full textInternational audienceAbstract Fused filament fabrication is among the most widely used 3D printing techniques, that allows to create complex devices from a continuous filament of a polymeric material. This adaptable technique has garnered considerable interest for the development of components functioning under severe radiation environments, including particle accelerators and nuclear reactors. In this work, we evaluated the radiation effects on several categories of commercial printing materials, namely the poly(lactic acid) (PLA), the acrylonitrile butadiene styrene (ABS), and a thermoplastic elastomer (TPE). Printing filaments have been exposed to x-rays (up to 160 keV) at 0.60 Gy s −1 , from 45 kGy to 2 MGy. The results from tensile tests, thermal analyses (differential scanning calorimetry, thermogravimetric analysis, dynamic mechanical thermal analysis), and spectroscopy tests (Fourier transform infrared spectroscopy and Raman analysis) reveal a dose-dependent degradation of material properties, predominantly affecting mechanical properties rather than chemical and thermal ones. The PLA shows the lowest radiation tolerance among the three, dramatically decreasing the tensile strength above 100 kGy, while TPE and ABS reach a comparable mechanical degradation after 1 MGy and 2 MGy, respectively. Radiation-induced effects are investigated, and the degradation is primarily attributed to chain scission as the principal damage mechanism
High-Sensitivity Microwave Detection of Subwavelength Objects Using a Dielectric-Loaded Waveguide with Electromagnetic Jet Lens
No full textInternational audienceThis letter presents a high-sensitivity, non-contact sensing approach based on a 15 GHz electromagnetic jet for the detection of subwavelength-scale objects. The jet is generated by a rectangular waveguide partially filled with polytetrafluoroethylene (PTFE) and terminated by a rectangular dielectric lens extending into free space. An impedancematching transition between the empty and PTFE-filled sections of the waveguide is optimized to minimize insertion loss resulting in a single, easily insertable dielectric component that ensures seamless integration into the waveguide. The lens geometry is designed through full-wave simulations to achieve subwavelength focusing. The proposed structure produces a focal spot with Full Width at Half Maximum (FWHM) of 0.41λ and 0.59λ in the x and y-directions, respectively. Experimental results confirm the enhanced near-field confinement and demonstrate the detection of objects as small as λ/400, highlighting advantages over conventional open-ended waveguide configurations. Owing to its fully guided architecture, compactness, and ease of use, the proposed device is particularly well suited for high-resolution, non-invasive sensing and offers strong potential for transfer to industrial applications
Premature baby cries remain different from those of full-term babies
No full textInternational audiencePrematurity is known to increase the risk of organic or functional pathologies from the neonatal period as well as after birth. The impact of prematurity on the acoustic structure of cries, which are vital signals for survival of babies, has received little scrutiny, despite its prognostic potential. Here, we test the hypothesis that the acoustic structure of the cries of preterm babies changes with their growth, ultimately becoming identical to the cries of full-term infants. We recorded and analyzed the cries of 18 premature babies, born between gestational age 26 and 33 weeks, every week while they were hospitalized in a neonatal intensive care unit. Our results show that as premature babies approach their theoretical term, their cries become increasingly tonal, less marked by noisy components, and carrying individual cues. We then compared these cries with those recorded in 25 healthy full-term babies. Despite the evolution of premature baby cries over time, the acoustic structure of these cries remains different from that of full-term infants. These results suggest that the vocal maturity of premature babies is not fully attained at the theoretical age of term
Using Signatures and Koopman Operator to Learn Nonlinear Dynamics
No full textInternational audienceWe propose a novel framework for predicting the evolution of dynamical systems by learning the Koopman operator in the space of linear functionals on the Signature transform of trajectory data. The Signature, a central object in rough path theory, provides a universal and compact representation of paths through iterated integrals, enabling linear models to approximate a wide class of non-linear functionals. By restricting observables to lie in the span of truncated Signatures, we construct a finite-dimensional approximation of the Koopman operator, which we estimate directly from data using regularized linear regression. This approach merges the expressiveness of operator-theoretic methods with the structural richness of Signature features.</div
Apprentissage incrémental de l'opérateur de Koopman pour systèmes non-autonomes via prédiction conforme
No full textInternational audienceThe Koopman operator enables the linearization of certain nonlinear dynamical systems by acting on observables, but its extension to the non-autonomous setting remains challenging. We propose an incremental method to estimate the Koopman operator in a non-autonomous context, leveraging conformal prediction to automatically detect regime changes. Model parameters are updated only when local performance deteriorates, ensuring sparse updates while maintaining accuracy comparable to off-line training. Preliminary experiments on the Duffing oscillator illustrate the effectiveness of our approach.L'opérateur de Koopman permet de linéariser certaines dynamiques non linéaires via une action sur les observables, mais son extension au cas non-autonome reste un défi. Nous proposons une méthode incrémentale pour estimer cet opérateur dans un cadre non-autonome, en s'appuyant sur la prédiction conforme pour détecter automatiquement les changements de régime. Les paramètres du modèle sont ajustés uniquement lorsque la performance locale se dégrade, assurant des mises à jour parcimonieuses tout en conservant une précision comparable à l'apprentissage hors-ligne. Des expériences préliminaires sur le système de Duffing illustrent l'efficacité de notre approche
Analysis of the Thermal Annealing Regeneration Process of RPL Dosimeters for High Dose Levels
No full textInternational audienc
Chaines d'atomes uniques de dysprosium dans des pinces optiques pour étudier la diffusion collective de la lumière
No full textThis thesis focuses on the development of an optical tweezer experiment with dysprosium atoms to study light-induced interactions.The goal is to investigate situations in which the electric field radiated by an atom as it de-excites can influence another nearby atom.The work presented here shows that the techniques developed to image and manipulate individual atoms in optical tweezers can provide new experimental tools to study the problem of collective light scattering.The first part of the thesis involves the design and construction of an optical tweezer experiment using dysprosium for the first time.Once the system was built, we characterized the behavior of dysprosium atoms in optical tweezers.In particular, we measured the influence of the trap light's polarization on the atom's polarizability.Through these measurements, we demonstrated the existence of a magic polarization that enables high-fidelity, non-destructive imaging of individual dysprosium atoms.Once this goal was achieved, we carried out some light scattering experiments on ordered atomic arrays created with optical tweezers.To do so, we developed a technique to detect the internal state of a single atom on an optical transition, leveraging the rich internal structure of the dysprosium atom.Using this technique, we studied how the electric field radiated by one atom can influence the measured frequency of the transition.We observed a collective shift of the spectrum in the steady-state regime, within the approximation of classical radiating dipoles.We then measured the influence of these interactions on Ramsey spectroscopy.This situation cannot be described within a classical approximation and requires to take into account the discrete nature of the energy levels of the atoms.We also establish a connection between the results obtained in this regime and the classical limit.Cette thèse porte sur la construction d'une expérience de pinces optiques avec des atomes de dysprosium pour étudier les interactions induites par la lumière.Le but est d'étudier des situations dans laquelle le champ électrique rayonné par un atome lorsqu'il se désexcite peut influencer un autre atome proche.Le travail présenté ici montre que les techniques développées pour imager et manipulerdes atomes individuels dans des pinces optiques peuvent fournir de nouveaux outils expérimentaux pour étudier le problème de la diffusion collective de la lumière.Une première partie de la thèse consiste en la conception et la construction d'une expérience de pinces optiques utilisant pour la première fois des atomes de dysprosium. Une fois le système construit, nous avons caractérisé le comportement des atomes de dysprosium dans les pinces optiques.En particulier, nous avons mesuré l'influence de la polarisation de la lumière du piège sur la polarisabilité de l'atome.Grâce à ces mesures, nous avons montré l'existence d'une polarisation magique qui rend possible l'imagerie non-destructive d'atomes uniques de dysprosium avec haute fidélité.Une fois ce but atteint, nous avons réalisé de premières expériences de diffusion de la lumière sur des ensembles atomiques ordonnés réalisés avec des pinces optiques.Pour ce faire, nous avons développé une technique de détection de l'état interne d'un atome unique sur une transition optique en utilisant la structure riche de l'atome de dysprosium.Grâce à cette technique, nous avons étudié comment le champ électrique rayonné par un atome sur un autre influence la mesure de la fréquence de la transition.Nous avons observé le décalage collectif du spectre en régime stationnaire dans l'approximation de dipôles rayonnants classiques.Nous avons ensuite mesuré l'influence de ces interactions sur la spectroscopie de Ramsey.Cette situation ne peut être décrite dans une approximation classique et il est nécessaire de tenir compte de la discrétisation de l'état interne des atomes.Nous faisons également le lien entre les résultats obtenus dans ce régime et la limiteclassique
T-matrix representation of optical scattering response: Suggestion for a data format
No full textInternational audienceThe transition matrix, frequently abbreviated as T-matrix, contains the complete information in a linear approximation of how a spatially localized object scatters an incident field. The T-matrix is used to study the scattering response of an isolated object and describes the optical response of complex photonic materials made from ensembles of individual objects. T-matrices of certain common structures, potentially, have been repeatedly calculated all over the world again and again. This is not necessary and constitutes a major challenge for various reasons. First, the resources spent on their computation represent an unsustainable financial and ecological burden. Second, with the onset of machine learning, data is the gold of our era, and it should be freely available to everybody to address novel scientific challenges. Finally, the possibility of reproducing simulations could tremendously improve if the considered T-matrices could be shared. To address these challenges, we found it important to agree on a common data format for T-matrices and to enable their collection from different sources and distribution. This document aims to develop the specifications for storing T-matrices and associated metadata. The specifications should allow maximum freedom to accommodate as many use cases as possible without introducing any ambiguity in the stored data. The common format will assist in setting up a public database of T-matrices
Enhancing the robustness of underwater dehazing by jointly using polarization and the dark channel prior
No full textInternational audienceThe scattering of light in water generates a veiling effect in underwater images, which significantly reduces visibility and limits the range of imaging systems, particularly in target detection applications. Several dehazing techniques have been developed based on the assumption that the veil and the scene of interest exhibit distinct physical properties, either colorimetric or polarimetric. However, these approaches tend to lose effectiveness when the underlying assumptions are not met. In this study, we propose a method that jointly exploits both the colorimetric and polarimetric characteristics of the veil to enhance image contrast. The proposed approach is evaluated on in situ images acquired using an underwater imaging system in both passive and active configurations. We demonstrate that our method enables the automatic identification of pixels associated with the veil, facilitating the estimation of its physical properties, and reduces potential ambiguities between the veil and regions of the scene of interest. Experimental results show that our approach achieves performance comparable to conventional techniques while offering improved robustness
Self-Play Q-Learners Can Provably Collude in the Iterated Prisoner's Dilemma
No full textInternational audienceA growing body of computational studies shows that simple machine learning agents converge to cooperative behaviors in social dilemmas, such as collusive price-setting in oligopoly markets, raising questions about what drives this outcome. In this work, we provide theoretical foundations for this phenomenon in the context of self-play multi-agent Q-learners in the iterated prisoner's dilemma. We characterize broad conditions under which such agents provably learn the cooperative Pavlov (win-stay, lose-shift) policy rather than the Pareto-dominated "always defect" policy. We validate our theoretical results through additional experiments, demonstrating their robustness across a broader class of deep learning algorithms