HAL-INSA Toulouse
Not a member yet
34325 research outputs found
Sort by
Wave propagation phenomena in nonlinear hierarchical neural networks with predictive coding feedback dynamics
No full textInternational audienceWe propose a mathematical framework to systematically explore the propagation properties of a class of continuous in time nonlinear neural network models comprising a hierarchy of processing areas, mutually connected according to the principles of predictive coding. We precisely determine the conditions under which upward propagation, downward propagation or even propagation failure can occur in both bi-infinite and semi-infinite idealizations of the model. We also study the long-time behavior of the system when either a fixed external input is constantly presented at the first layer of the network or when this external input consists in the presentation of constant input with large amplitude for a fixed time window followed by a reset to a down state of the network for all later times. In both cases, we numerically demonstrate the existence of threshold behavior for the amplitude of the external input characterizing whether or not a full propagation within the network can occur. Our theoretical results are consistent with predictive coding theories and allow us to identify regions of parameters that could be associated with dysfunctional perceptions
Does Electronic Strong Light-Matter Coupling Affect the Ground-State Energy Landscape? An Experimental Study Using Spin-Crossover Molecules
No full textInternational audienceThe effect of strong light-matter coupling on the electronic ground-state energy landscape of a large ensemble of coupled molecules remains an open question, even at the theoretical level, which still suffers from the lack of experimental studies. In the present work, we have conducted a very careful study of the thermodynamic phase equilibrium between the low-spin (LS) and high-spin (HS) states of a molecular spin-crossover (SCO) thin film, strongly coupled to the vacuum field inside a Fabry–Pérot cavity. While the cavity was tuned to be resonant with the intense charge-transfer bands of the SCO complexes in the LS state, allowing a strong-coupling regime to be achieved with a Rabi splitting of up to 670 meV, molecules in the nonabsorbing HS state remain uncoupled to the cavity. Importantly, no significant change in the spin-transition temperature is observed between the LS and HS states under light-matter coupling within the precision limit (1 °C) of our measurements. The present results demonstrate that, although collective strong coupling to electronic excitations can significantly perturb the excited states of molecules, the effect on the ground-state energy levels remains largely negligible (<0.6 meV)
H 2 S Sensing with SnO 2 ‐Based Gas Sensors: Sulfur Poisoning Mechanism Revealed by Operando DRIFTS and DFT Calculations
No full textInternational audienceAbstract Real‐time detection of toxic and flammable H 2 S remains challenging for cost‐effective semiconducting metal oxide (SMOX) sensors due to the insufficient focus on and inherently poor understanding of the sulfur‐poisoning effect. This research, focusing on SnO 2 as a model for SMOX sensors, identifies the formation of sticky sulfite and sulfate surface species as the root cause of poisoning through the detailed analyses of results obtained from operando diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments and density functional theory (DFT) calculations. The formation of the poisoning species is highly energetically favorable. Meanwhile, the decomposition of sulfite and sulfate appears unfavorable at the typical operating temperature of 300 °C and is only feasible around the literature‐reported 500 °C. The sulfur poisoning effect is also likely to occur with SO 2 and other sulfur‐containing volatile organic compounds (VOCs). Overcoming this issue is expected to require surface additives and/or alternative SMOX materials capable of providing different reaction pathways. The significance of metal‐sulfur‐oxygen chemistry extends beyond SMOX gas sensors to desulfurization catalysts, denitration catalysts, and solid oxide fuel cells
Gaussian quantum information over general quantum kinematical systems I: gaussian states
No full textInternational audienceWe develop a theory of Gaussian states over general quantum kinematical systems with finitely many degrees of freedom. The underlying phase space is described by a locally compact abelian (LCA) group G with a symplectic structure determined by a 2-cocycle on G. We use the concept of Gaussian distributions on LCA groups in the sense of Bernstein to define Gaussian states and completely characterize Gaussian states over 2-regular LCA groups of the form endowed with a canonical normalized 2-cocycle. This covers, in particular, the case of n-bosonic modes, n-qudit systems with odd , and p-adic quantum systems. Our characterization reveals a topological obstruction to Gaussian state entanglement when we decompose the quantum kinematical system into the Euclidean part and the remaining part (whose phase space admits a compact open subgroup). We then generalize the discrete Hudson theorem (Gross in J Math Phys 47(12):122107, 2006) to the case of totally disconnected 2-regular LCA groups. We also examine angle-number systems with phase space and fermionic/hard-core bosonic systems with phase space (which are not 2-regular) and completely characterize their Gaussian states
Nucleation in Confinement: A Pathway to Size-Controlled Ultra-Small Icosahedral Silver Nanoparticles
No full textInternational audienceThe size and structural control of ultra-small metal nanoparticles (NPs) remains a key area of research, as their unique nanoscale properties can significantly impact their utility and applications in industry. Among the different synthesis strategies, the formation of large clusters of metal precursors in solution, which favours a confined nucleation, has been less explored. Here, we report a simple synthesis protocol of ultra-small silver nanoparticles with an icosahedral structure, a mean diameter of 3 nm and low polydispersity (down to 6%), through the reduction of AgNO3 by trialkylsilanes in a solution of oleylamine (OY) and hexane. Smalland wide-angle X-ray scattering (SAXS, WAXS) and pair distribution function (PDF) analysis reveal that the Ag NP size and structure remain constant regardless of Ag and reducing agent concentrations, or temperatures up to 120°C. The lack of correlation between the final NP size and the initial Ag concentration, temperature, and reaction rate -despite variations spanning two orders of magnitude -does not align with a classical nucleation/growth mechanism of isolated monomers in solution. SAXS analysis of the precursor solutions revealed the presence of {OY-Ag (I) -NO3}n clusters, which interact repulsively in solution. The final Ag particle size is closely tied to the size of the Ag(I) clusters. This suggests that size control is governed by a nucleation process involving Ag atoms within the clusters. The separation of nucleation and growth appears to be achievable through nucleation confined within the repulsively interacting {OY-Ag (I) -NO3}n clusters. Only two factors have been found to decrease the mean NP size: increasing the OY concentration, or substituting AgNO3 with silver trifluoro-sulfonate or trifluoro-acetate, which reduces the mean size, in agreement with smaller initial Ag (I) clusters. The advantage of this synthesis route is the very high concentration of metals that can be achieved without degrading the morphological and structural quality of the final particles
Conceptual Approach for Aerobic Autotrophic Gas Cultivation in Shake Flasks: Overcoming the Inhibitory Effects of Oxygen in Cupriavidus necator
No full textInternational audienceThis study conceptualizes the design of a small‐scale system (250 mL–1 L) for the autotrophic cultivation of hydrogen‐oxidizing bacteria, such as the representative strain Cupriavidus necator . The research aimed to systematically investigate the impact of bottle volume and gas composition, particularly oxygen concentration, on the growth and performance of C. necator during autotrophic cultivations. To this end, customized, pressure‐tight, baffled glass bottles of various sizes (250, 500, and 1000 mL) and gas mixtures with varying oxygen concentrations (4%, 8%, and 12% v/v) were tested. Growth was monitored by measuring optical density. The maximum specific growth rate ( µ max ), the biomass production rate (BPR), the volumetric gas–liquid mass transfer coefficient ( k L a ), and the oxygen transfer rate were calculated. Among the various combinations, the 1000‐mL bottles demonstrated the highest µ max (0.13 h −1 ) and the second‐highest BPR (0.074 g L −1 h −1 ) at an oxygen concentration of 8%, without the need to refill the headspace. The proposed small‐scale system offers a swift and replicable method for concurrently investigating multiple autotrophic cultivations. In this regard, increasing the size of the bottle flask proved to be an efficient strategy to minimize the periodicity for gas refilling. Due to the inhibitory effect of oxygen, changing the liquid–gas volume ratio in hydrogen‐driven shake flask cultivation had so far strongly influenced the growth rate. Our results provide a solid foundation for the scaling and optimization of small‐scale cultivation of chemolithotrophic bacteria and will facilitate future parallelization and, hence, optimization of metabolic aspects
Collision Avoidance in Model Predictive Control using Velocity Damper
No full textWe propose an advanced method for controlling the motion of a manipulator robot with strict collision avoidance in dynamic environments, leveraging a velocity damper constraint. Unlike conventional distance-based constraints, which tend to saturate near obstacles to reach optimality, the velocity damper constraint considers both distance and relative velocity, ensuring a safer separation. This constraint is incorporated into a model predictive control framework and enforced as a hard constraint through analytical derivatives supplied to the numerical solver. The approach has been fully implemented on a Franka Emika Panda robot and validated through experimental trials, demonstrating effective collision avoidance during dynamic tasks and robustness to unmodeled disturbances. An efficient open-source implementation along examples are provided here: https://gepettoweb.laas.fr/articles/ haffemayer2025.html
Tree-Based Diagnosis Enhanced with Meta-Knowledge and Discovery of Analytical Redundancy Relations
No full textInternational audienc
FISTA restart using an automatic estimation of the growth parameter
No full textInternational audienceIn this paper, we propose a restart scheme for FISTA (Fast Iterative Shrinking-Threshold Algorithm). This method which is a generalization of Nesterov's accelerated gradient algorithm is widely used in the field of large convex optimization problems and it provides fast convergence results under a strong convexity assumption. These convergence rates can be extended for weaker hypotheses such as the \L{}ojasiewicz property but it requires prior knowledge on the function of interest. In particular, most of the schemes providing a fast convergence for non-strongly convex functions satisfying a quadratic growth condition involve the growth parameter which is generally not known. Recent works show that restarting FISTA could ensure a fast convergence for this class of functions without requiring any knowledge on the growth parameter. We improve these restart schemes by providing a better asymptotical convergence rate and by requiring a lower computation cost. We present numerical results emphasizing the efficiency of this method
A model for framed configuration spaces of points
No full text47 pages, revised and expanded version. Comments are welcomeWe study configuration spaces of framed points on compact manifolds. Such configuration spaces admit natural actions of the framed little discs operads, that play an important role in the study of embedding spaces of manifolds and in factorization homology. We construct real combinatorial models for these operadic modules, for compact smooth manifolds without boundary