Université de technologie de Troyes open archive
Not a member yet
10722 research outputs found
Sort by
Limiting empirical spectral measure of the normalized Laplacian in preferential attachment graphs: Normalized Laplacian spectrum in preferential attachment
No full textInternational audienceWe study the empirical spectral distribution of the normalized Laplacian of linear preferential attachment graphs in the Barabási-Albert regime with fixed out-degree. For the resulting sequence of random multigraphs, we prove that the empirical spectral distribution converges weakly in probability to a deterministic probability measure supported on the interval [0, 2]. The limit is characterized via the local weak limit of preferential attachment graphs (the Pólya-point graph): the limiting Stieltjes transform is given by the expected diagonal Green function at the root of the normalized Laplacian operator on this infinite random graph. The proof combines a resolvent approach with a uniform Neumann-series expansion for the normalized Laplacian, a random-walk representation in terms of return probabilities on decorated neighborhoods, a truncation and Doob martingale-Azuma-Hoeffding concentration argument along the PA filtration, and an analytic continuation argument based on normal families.Nous étudions la distribution spectrale empirique du Laplacien normalisé des graphes d'attachement préférentiel linéaire dans le régime de Barabási-Albert, avec degré sortant fixé. Pour la suite de multigraphes aléatoires ainsi obtenue, nous montrons que la distribution spectrale empirique converge faiblement en probabilité vers une mesure de probabilité déterministe supportée sur l'intervalle [0,2].La limite est caractérisée à partir de la limite locale faible des graphes d'attachement préférentiel (le graphe à points de Pólya) : la transformée de Stieltjes limite est donnée par l'espérance de la fonction de Green diagonale à la racine de l'opérateur du Laplacien normalisé sur ce graphe aléatoire infini.La preuve combine une approche par résolvante avec un développement uniforme en série de Neumann pour le Laplacien normalisé, une représentation en marche aléatoire en termes de probabilités de retour sur des voisinages décorés ainsi qu'un argument de troncature et de concentration de type martingale de Doob–Azuma–Hoeffding le long de la filtration d'attachement préférentiel
Unified Anomaly Detection via Multi-Scale Contrasted Memory
No full textInternational audienceDeep anomaly detection aims to provide robust and efficient classifiers for zero-shot (unsupervised, UNS) and few-shot (imbalanced supervised, IMS) settings. However, current models still struggle on edge-case normal samples and are often unable to keep high performance over different scales of anomalies. Additionally, there is a lack of a unified framework that efficiently addresses both UNS and IMS settings. To address these limitations, we present a novel two-stage method which leverages multi-scale normal prototypes during training to compute an anomaly deviation score. First, we employ a novel memory-augmented contrastive learning to jointly learn representations and memory modules across multiple scales. This allows us to effectively capture subtle features of normal data while adapting to varying levels of anomaly complexity. Then, we train an efficient anomaly distance-based detector that computes spatial deviation maps between the learned prototypes and incoming observations. Our model outperforms the SoTA on a wide range of anomalies, including object, style, and local anomalies, as well as industrial inspection and face anti-spoofing, while being on par with SoTa out-of-distribution detectors. Notably, it stands as the first model capable of maintaining exceptional performance across both settings
Patients visits forecasting in emergency departments: Self-adapting LSTM models for evolving data distributions
No full textInternational audienc
Conflict-Free Path Planning and Scheduling for Rgvs in Automated Warehousing Systems
No full textInternational audienc
Recovery of lithium, nickel and cobalt from mixed lithium-ion batteries in ammonia media
No full textInternational audienceThis study investigates the recovery of valuable metals from a mixture of lithium iron phosphate (LFP) and nickel-manganese-cobalt (NMC) in an ammoniacal medium. A factorial design of experiments was applied to assess the effects of NH 3 , (NH 4 ) 2 SO 4 , and Na 2 SO 3 concentrations on the leaching behavior of Ni, Mn, Co, and Li. Ammonia enhanced metal dissolution through complexation, (NH 4 ) 2 SO 4 stabilized the pH via buffering, and Na 2 SO 3 acted as a reducing agent improving the dissolution of Ni, Co, and Li. Optimal conditions (6 mol L -1 NH 3 , 3 mol L -1 (NH 4 ) 2 SO 4 , 1 mol L -1 Na 2 SO 3 , 60 • C, S/L = 50 g L -1 ) yielded recoveries of 68.1% Ni, 25.9% Mn, 79.4% Co, 55.6% Li, and 99.7% Cu. An empirical model accurately predicted these dissolution yields. Subsequent solvent extraction enabled selective recovery of Ni, Mn, and Cu using Acorga® M5640, and Li using Cyanex® 936P diluted in kerosene. The proposed flowsheet demonstrates the efficiency and selectivity of ammoniacal leaching as a sustainable alternative to conventional acid-based recycling of lithium-ion batteries
Influence of electrical properties on thermal boundary conductance at metal/semiconductor interface
No full textInternational audienceRecent experimental investigations have demonstrated that doping a semiconductor is a route to increase the thermal boundary conductance at metal/semiconductor interfaces. In this work, the influence of the electrical properties on heat transfer across metal/doped semiconductor junctions is investigated. Specifically, thermal boundary conductance at the interfaces between p- and n-doped silicon and titanium is measured by employing frequency-domain photothermal radiometry under varying external conditions. The influence of the doping level of the semiconductor, the barrier height, and the space charge area is analyzed. In particular, a 40% increase in the interface thermal conductance with the application of a current at n-doped silicon/titanium interfaces is reported. The enhancement of the thermal boundary conductance is explained by the shrinking of the surface charga area induced by the electric current. This study opens the way to modulating interfacial heat transfer at metal/semiconductor interfaces through fine tuning of electrical effects
Multilayered hyperbolic Au/TiO 2 nanostructures for enhancing the nonlinear response around the epsilon-near-zero point
No full textInternational audienceIn this work, we present the design, fabrication, and study of the optical properties of multilayered metal–dielectric Au/TiO 2 structures. The samples were fabricated using Joule effect evaporation for gold and electron beam evaporation for titanium dioxide. Their structure was designed to have an epsilon-near-zero (ENZ) point at different wavelengths around 800 nm, in order to study their nonlinear response as a function of the resonance conditions around the ENZ point. The characterization of the linear properties of the samples was done using spectrophotometry and spectral ellipsometry. We studied the nonlinear response with the z-scan technique at different incident irradiances using a Ti:sapphire femtosecond laser, enabling us to characterize both the refractive and absorptive contributions to the nonlinear response. Due to the high pulse repetition rate inherent to Ti:sapphire systems and the presence of linear absorption in the samples, cumulative pulse-to-pulse thermal effects may be present. A modified version of the z-scan technique that allowed us to separate the electronic from the thermal contribution was used. A clear enhancement of the nonlinear response was observed for the sample with an ENZ point around the laser wavelength 800 nm with a nonlinear refractive index of n 2 = 0.103 ± 0.006 cm 2 ·GW −1 , a value that is comparable to other ENZ materials in literature
Joint Maintenance and Routing under Sudden Shocks: Modeling and Sensitivity Analysis
No full textInternational audienceJoint Maintenance and Routing under Sudden Shocks: Modeling and Sensitivity Analysi
Towards a Movement-sensitive Approach for Collaborative Digital Transformation in Ecologies of Artifacts
No full textToday's digital transformation projects are scarcely beginning from a blank page as they should deal with an installed based of artifacts and practices. The concept of Ecology of Artifacts (EoA) provides an interesting framework highlighting how collaborative activities span over multiple tools that belongs to the participants and their organizations. EoAs are dynamic and undergo evolution over time. These evolutions can be accounted for as movements where the introduction, the replacement or decommissioning of a tool can impact practices and other artifacts, like ripple effects. This exploratory paper presents an under development workshop-based method that aims at supporting collaborative reflection about EoA evolutions. The workshop is structured into two phases: the description of the practice's EoA and the collaborative analysis of its transformations with the participants. We report about a formative expert evaluation of our method with a group of academics engaged in hybrid meetings interested with changing their tools. The participants feedback indicates that the approach is useful for describing the EoA of a collaborative practice and for analyzing movements within them. Their feedback also suggest improvements for future development of our movement-sensitive approach
Assessing Quantum Annealing to solve the Minimum vertex multicut
No full textInternational audienceCybersecurity in telecommunication networks often leads to hard combinatorial optimization problems that are challenging to solve with classical methods. This work investigates the practical feasibility of using quantum annealing to address the Restricted Vertex Minimum Multicut Problem. The problem is formulated as a Quadratic Unconstrained Binary Optimization model and implemented on D-Wave’s quantum annealer. Rather than focusing on solution quality alone, we analyze key aspects of the quantum workflow including minor embedding techniques, chain length, topology constraints, chain strength selection, unembedding procedures, and postprocessing. Our results show that quantum annealing faces substantial hardware-level constraints limitations in embedding and scalability, especially for large in-stances, while hybrid quantum-classical solvers provide improved feasibility. This study offers a realistic assessment of the D-Wave system’s current capabilities and identifies crucial parameters that govern the success of quantum optimization in cybersecurity-related network problems