HAL Portal IOGS (nstitut d'Optique Graduate School)
Not a member yet
12589 research outputs found
Sort by
On Jitter Transfer in Ring Oscillators and Comprehensive Modelling of 1/f Noises
No full textInternational audienceRing oscillator jitter serves as a crucial entropy source for provably secure True Random Number Generators (TRNGs). This talk addresses two gaps in the existing research. First, practical implementations use coupled oscillators (one serving as a reference clock), and their joint stochastic dynamics remain only approximately understood, potentially leading to inaccuracies in security assessment. Second, numerical challenges in quantitative modeling of low-frequency noises make the entropy source underutilized.The first part of this talk presents an analytical solution for the relative stochastic dynamics of coupled ring oscillators, providing formal justification for the jitter transfer principle—a heuristic approach where one oscillator is assumed jitter-free while the other is jitter-compensated. These insights enable more accurate jitter estimation in multi-ring oscillator systems, improving TRNG performance (joint work with David Lubicz).The second part presents a general and scalable framework for modeling low-frequency noises with negative power law, through Fractional Brownian Motion and forecasting properties of Gaussian Processes, inspired by the pioneering work of atomic clock physicists D.W. Allan and J.A. Barnes
Développement de revêtements interférentiels pour des imageurs X à haute résolution
No full textInertial Confinement Fusion (ICF) is a preferred experimental approach to access extreme matter conditions, through the implosion of a laser-driven target. To characterize the implosion symmetry, a micrometer-resolution microscope, operating in the hard X-ray range, is being developed by the CEA (Commissariat à l'énergie atomique). TXI (Toroidal X-ray Imager), which will be installed at the NIF (National Ignition Facility), is a Wolter-type X-ray diagnostic where conical mirrors are replaced by toroidal mirrors. It is also a multi-channel diagnostic, operating at a nominal grazing angle of 0.6°, allowing imaging at 8.7 keV, 13 keV, and 17.5 keV. The required thicknesses of the multilayer coatings must become increasingly thin to image these energies. Different multilayer formulas (alternating two materials whose total period allows reflection of a certain wavelength, according to Bragg's law) have been optimized to meet TXI's specifications. The instrument's optical response was simulated using ray-tracing software. The coatings were then produced by sputtering deposition. For the next phase of the thesis, a preliminary study was conducted on designing an imager capable of operating up to 60 keV, as well as a pre-study on HiPIMS (High Power Impulse Magnetron Sputtering) technology to assess its benefits for thin-film quality.La FCI (fusion par confinement inertiel) est une voie privilégiée pour accéder expérimentalement aux conditions extrêmes de la matière, via l'implosion d'une cible par laser. Pour caractériser la symétrie d'implosion, un microscope de résolution micrométrique, opérant dans le domaine des rayons X durs, est développé par le CEA (Commissariat à l'énergie atomique). TXI (Toroidal X-ray Imager) qui sera installé au NIF (National Ignition Facility) est un diagnostic X de type Wolter, où les miroirs coniques sont remplacés par des miroirs toriques. Il est également un diagnostic multicanal, fonctionnant à un angle de rasance nominal de 0.6°, et permettant d'imager des énergies de 8.7 keV, 13 keV et 17.5 keV. Les épaisseurs requises de revêtements multicouches doivent être de plus en plus fines pour imager ces énergies. Différentes formules de multicouches (alternances de deux matériaux dont la période totale permet de réfléchir une certaine longueur d'onde, conformément à la loi de Bragg) ont été optimisés, afin de satisfaire le cahier des charges de TXI. La réponse optique de l'instrument a été simulée à l'aide d'un logiciel de tracé de rayon. Les revêtements ont ensuite été réalisés par pulvérisation cathodique. Pour la suite de la thèse, une pré-étude sur la conception d'un imageur fonctionnant jusqu'à 60 keV a été menée, ainsi qu'une pré-étude sur la technologie HiPIMS (High Power Impulse Magnetron Sputtering) pour en étudier les bénéfices sur la qualité des films minces
High-confidence Remote Power Analysis on Heterogeneous SoCs
No full textInternational audienceIn recent years, significant advances have been made in side-channel analysis, particularly in the design of attack methodologies targeting SoC-FPGAs. These devices have become increasingly popular in cloud data centers thanks to their flexibility and efficiency. As a result, there has been a growing number of proposals for sharing FPGA fabrics among multiple users in cloud environments. However, even when logical isolation is used as a protective measure for each tenant, the presence of a multiple-tenants in the FPGA environment raises significant concerns about potential security threats. Recent works have revealed the possibility of power side-channel attacks being executed in a cloud-FPGA environment, even without direct physical access to the platform. These attacks can be carried out by a malicious user, leveraging delay sensors implemented using internal FPGA resources. These sensors have the ability to monitor the power consumption of a circuit, thus giving a malicious user insights into the internal operations of the SoC-FPGA and potentially enabling extraction of sensitive information. The primary challenge for successful remote power analysis lies in accurately cutting and aligning power traces. Secondary digital channels with trigger information are typically used for this purpose. This paper presents a novel method that simplifies the conditions necessary for a remote power attack. The approach mitigates the need to connect digital triggers to the remote sensor, thereby reducing the complexity of the attack setup. To validate the efficacy of the proposed method, a successful key recovery was performed on a hardware implementation of an AES cipher.\end{abstract
Stabilization of the spectral power distribution of a tunable multichannel LED lighting system
No full textInternational audienceThe advancements in Light-Emitting Diodes (LEDs) have allowed spectrally tunable light sources to gain attention in many fields of research thanks to their ability to produce a specific light output. However, LED outputs can fluctuate with temperature, and aging components can lead to noticeable discrepancies in light characteristics. This study thoroughly examines the Telelumen Dittosizer light player LED panel to exemplify a commercially available device and the associated challenges in predicting and stabilizing its output. Then, we introduce an innovative algorithm aimed at addressing such a stabilization challenge, based on a straightforward characterization procedure along with an external spectrometer. The accuracy of the algorithm was validated with different inputs, achieving a ΔE,2000 lower than 0.5. Our findings demonstrate the ability to stabilize the spectral power distribution for a minimum of 30 min. The proposed algorithm is hardware-independent and adaptable to any combination of spectrally tunable light sources and spectrometers
Linear Modeling of the Adversarial Noise Space
No full textInternational audienceRecent works have revealed the vulnerability of deep neural network (DNN) classifiers to adversarial attacks. Among such attacks, it is common to distinguish specific attacks adapted to each example from universal ones, referred to as example-agnostic. Even though specific adversarial attacks are efficient on their target DNN classifier to attack, they struggle to transfer to others. Conversely, universal adversarial attacks suffer from lower attack success. To reconcile universality and efficiency, we propose LIMANS, a model of the adversarial noise space, allowing to frame any specific adversarial perturbation as a linear combination of the universal adversarial directions. We bring in two stochastic gradient based algorithms for learning these universal directions and the associated adversarial attacks. Empirical analyses conducted with the CIFAR-10 and ImageNet datasets show that LIMANS (i) enables crafting specific and robust adversarial attacks with high probability, (ii) provides a deeper understanding of DNN flaws, and (iii) shows significant ability in transferability
Enhanced Spontaneous Light Emission of ZnO Nanowire-Based Gratings
No full textInternational audienceWe theoretically and experimentally investigated the spontaneous light emission of ZnO nanowires (NWs) arranged in grating structures. Through systematic calculations of light extraction and absorption efficiencies together with the Purcell factor, we shed light on the influence of disorder on these optical processes. This analysis unveils the role of optical resonances to achieve direct light extraction and to enhance the spontaneous emission. We demonstrate that the periodic structuring of the ZnO NWs at the micrometer scale is a robust and efficient solution to increase the spontaneous emission efficienc
Variational Perspective on Fair Edge Prediction
No full textInternational audienceAlgorithmic fairness has been of great interest in the machine learning community and more recently in the graph context. In this paper, we address the problem of dyadic fairness where the task at hand is edge prediction, and the population of interest (nodes) is divided into a protected and a non-protected group, e.g. men and women. The goal is then to ensure that there should be no statistically significant difference in the prediction outcomes between the two groups, after accounting for any relevant factors that may impact the outcome. To proceed, we design a novel loss based on the variational information bottleneck principle to learn individual node representation while controlling a given level of dyadic fairness. The optimization of the loss is done with a Graph Neural Network. Experiments carried out on several real-world datasets confirmed the capacity of the proposed method, to maintain high accuracy on the edge prediction task while significantly reducing potential bias
Local linear convergence of proximal coordinate descent algorithm
No full textInternational audienceFor composite nonsmooth optimization problems, which are "regular enough", proximal gradient descent achieves model identification after a finite number of iterations. For instance, for the Lasso, this implies that the iterates of proximal gradient descent identify the non-zeros coefficients after a finite number of steps. The identification property has been shown for various optimization algorithms, such as accelerated gradient descent, Douglas-Rachford or variance-reduced algorithms, however, results concerning coordinate descent are scarcer. Identification properties often rely on the framework of "partial smoothness", which is a powerful but technical tool. In this work, we show that partial smooth functions have a simple characterization when the nonsmooth penalty is separable. In this simplified framework, we prove that cyclic coordinate descent achieves model identification in finite time, which leads to explicit local linear convergence rates for coordinate descent. Extensive experiments on various estimators and on real datasets demonstrate that these rates match well empirical results
Wave function network description and Kolmogorov complexity of quantum many-body systems
No full textInternational audienceProgrammable quantum devices are now able to probe wave functions at unprecedented levels. This is based on the ability to project the many-body state of atom and qubit arrays onto a measurement basis which produces snapshots of the system wave function. Extracting and processing information from such observations remains, however, an open quest. One often resorts to analyzing low-order correlation functions - i.e., discarding most of the available information content. Here, we introduce wave function networks - a mathematical framework to describe wave function snapshots based on network theory. For many-body systems, these networks can become scale free - a mathematical structure that has found tremendous success in a broad set of fields, ranging from biology to epidemics to internet science. We demonstrate the potential of applying these techniques to quantum science by introducing protocols to extract the Kolmogorov complexity corresponding to the output of a quantum simulator, and implementing tools for fully scalable cross-platform certification based on similarity tests between networks. We demonstrate the emergence of scale-free networks analyzing data from Rydberg quantum simulators manipulating up to 100 atoms. We illustrate how, upon crossing a phase transition, the system complexity decreases while correlation length increases - a direct signature of build up of universal behavior in data space. Comparing experiments with numerical simulations, we achieve cross-certification at the wave-function level up to timescales of 4 s with a confidence level of 90%, and determine experimental calibration intervals with unprecedented accuracy. Our framework is generically applicable to the output of quantum computers and simulators with in situ access to the system wave function, and requires probing accuracy and repetition rates accessible to most currently available platforms