Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
Neural Network-Based Estimation of Optical Feedback Factor in Self-Mixing Interferometry Systems across various feedback regimes
No full textInternational audienceSelf-mixing interferometry (SMI) system is a promising and low-cost approach for displacement measurements. The resolution and precision of displacement estimation depend on key parameters of the SMI system and, in particular, on the optical feedback factor (C). This article presents a new method for estimating this C factor based on an artificial neural network (ANN) that analyzes certain specific fringe shape features of SMI signals. Specifically, only six features of the SMI signal are taken into account as input to a compact neural network with one hidden neuron layer (14 neurons) by the estimation process, which reduces the computational cost and simplifies its deployment. This study evaluates the C-estimation capability of the method in various simulated and experimental scenarios, including variations in displacement amplitude and frequency, random displacements, noise levels, and speckle effects. These analyses show that the method works correctly for a moderate feedback regime ( 14.6 ), this error is less than 2% (for C<8 ). Comparisons with other methods show similar or better accuracy but with the advantage of being able to handle all displacement waveforms (sinusoidal or arbitrary) at low computational cost, which facilitates integration into a real-time embedded system
Étude comparative de méthodes de Classification Hiérarchique d'images
No full textInternational audienceÉtude comparative de méthodes de Classification Hiérarchique d'image
Hacia una alfabetización crítica de la IA en la enseñanza de idiomas: antropomorfización, descolonización del conocimiento y agencia distribuida en contexto plurilingüe
No full textInternational audienc
Black Atlantic Space-Time Portfolio by Unity
No full textAt the Manifest exhibition in Nantes 2024.09.19 – 10.06 under the artistic name Unity, Deirdre C. Molloy and Aurélie Capelle-Sigère presented an audio-visual series conceived and produced by Molloy called Black Atlantic Space-Time. This illustrated portfolio PDF presents Unity's modular art exhibition of screendance, multimedia ethnography and an interactive map. The artworks and documentation are part of Molloy's PhD titled "From Blues Dance to Black Atlantic Identity: a postcolonial perspective". The portfolio documents innovative research-creation methods, collaborations and an original data-curation platform for intangible heritage. A scientific premise is rhythm code transmission through muscle memory rather than through DNA or 'blood'.RELATED MEDIABlack Atlantic Space-Time is expressed in 2 formats - a narrative film Drum Calls | Body Recalls and the interactive Unity Atlantic Map. Drum Calls Body Recalls is a poetic tribute to African ancestors and a resilient heritage of rhythm codes. Strong wifi and speakers or good headphones are recommended as the story is carried by the sound. Watch the Black Atlantic Space-Time teaser : https://doi.org/10.34847/NKL.5C091TX1To map Africanist intangible heritage, Molloy has made 16 short films to date, augmented by links to public and private archives and social media. Gerador provided software development for the map, informed by a prototype Molloy developed with Code Your Future UK students. The interactive map offers audio-visual and written context for rhythm codes and their circulations around the Atlantic. To hear/see the rhythms, click on the pink arrows and then click on 'Unmute'. https://unityatlantic.org/The narrative movie (19 mins) Drum Calls | Body Recalls is a poetic tribute to African ancestors through the diasporic heritage of rhythm codes. In order to arrange a screening of Drum Calls please contact the author. Here's a synopsis of Drum Calls: https://filmfreeway.com/DrumCallsBodyRecalls#ff_javascriptThis PDF portfolio documents the creative process of Black Atlantic Space-Time, including an AV/VR/XR residency series alongside 12 art projects (22 artists in total). The multimedia art and the group exhibition in Nantes 2024 were co-produced and promoted by 5 European cultural agencies, co-funded by the Creative Europe Program: Project Manifes
E-MATRO: An Adaptive and Energy-Efficient Framework for Slicing within the O-RAN Architecture
No full textInternational audienceThe complexity of Open RAN (O-RAN) networks, driven by dynamic traffic and stringent Quality of Service (QoS) requirements, challenges traditional centralized control, leading to suboptimal resource allocation and high energy consumption. We address the problem of energy-efficient, slice-compliant Virtual Network Function (VNF) placement by formulating it as a flow-based optimization model. To solve it, we propose E-MATRO, a multi-agent reinforcement learning (MARL) framework that integrates decentralized dApps with a centralized xApp for adaptive and scalable control. Experiments on simulated O-Cloud networks show that E-MATRO efficiently learns optimal policies, minimizing energy consumption while ensuring QoS compliance. Among tested RL approaches, Deep Q-Learning demonstrates superior adaptability in complex scenarios, highlighting the potential of hybrid MARL-based control for O-RAN
Le Ring à Paris (1933): chronologie et dossier de presse des quatre représentations, dezede.org, https://dezede.org/dossiers/id/935/
No full texthttps://dezede.org/dossiers/id/935
Structured Gaussian Process Regression for Multiphysics Problems
No full textInternational audienceIn addition to other multi-query problems, global optimization of design variables in multi-physics engineering problems requires the resolution of coupled systems of equations for many sets of design variables. Such coupled systems are often resolved using the partitioned approach, whereby each individual system of equations is solved separately using its numerical scheme, and the coupling equations are enforced through fixed point iteration. This optimization procedure can be made more efficient by using, for example, Gaussian Process based surrogate models. Gaussian Processes are attractive for this purpose, since they are flexible, have a strong theoretical foundation and the variance of the Gaussian Processes can be used to estimate the uncertainty of the surrogate solution. This estimate can be used to improve the accuracy of the surrogate model in areas of the parameter space where it is inaccurate. Previous work has explored the potential of applying this procedure to each distinct system of equations separately, and calculating the distribution of the solutions of the resulting coupled system through fixed point iteration of the sample paths of the Gaussian Processes. This algorithm was successfully used in Bayesian and gradient-based optimization. Conversely, the Gaussian Process modelling framework has recently been extended to handle various types of constraints. Novel strategies have been developed to ensure that Gaussian Process sample paths be, for example, monotonic, or convex. Constraining Gaussian Process sample paths to satisfy a given PDE has also been a popular line of inquiry. The aim of the present work is to extend these results concerning structured Gaussian Processes to the multiphysics case. First of all, ways for increasing the regularity of the Gaussian Process sample paths are investigated, with the objective of ensuring that the sample paths of the Gaussian Processes satisfy certain conditions that the solutions of the PDEs are known to satisfy. Certain strategies from the physics-constrained Gaussian Process literature are extended to the multiphysics case. Finally, we present some theoretical results concerning our work, and discuss and compare our work with others in the literature
"En guise de conclusion: la quête sans fin de Senancour"
No full textInternational audienceS'il est vrai que l'on trouve dans "Oberman" un discours de la fin, et notamment une représentation de sa propre fin par le rédacteur des lettres sous la forme d'un engourdissement, d'une usure des sens, d'un détachement, il faut aussi être sensible à l'élan qui se dessine dans le regard critique qu'il porte sur le monde et dans sa volonté de ne jamais conclure. Persistant, l'ennui laisse place à une joie intellectuelle et sensible dans les derniers moments des personnages, nimbés de douceur et de tendresse
Box-constrained L0 Bregman relaxations
No full textInternational audienceRegularization using the L0 pseudo-norm is a common approach to promote sparsity, with widespread applications in machine learning and signal processing. However, solving such problems is known to be NP-hard. Recently, the L0 Bregman relaxation (B-rex) has been introduced as a continuous, non-convex approximation of the L0 pseudo-norm. Replacing the L0 term with B-rex leads to exact continuous relaxations that preserve the global optimum while simplifying the optimization landscape, making non-convex problems more tractable for algorithmic approaches. In this paper, we focus on box-constrained exact continuous Bregman relaxations of L0-regularized criteria with general data terms, including least-squares, logistic regression, and Kullback-Leibler fidelities. Experimental results on synthetic data, compared with Branch-and-Bound methods, demonstrate the effectiveness of the proposed relaxations