Portail HAL Lumière Lyon 2
Not a member yet
111406 research outputs found
Sort by
BSP-OT: Sparse transport plans between discrete measures in loglinear time
No full textInternational audienceTo solve the optimal transport problem between two uniform discrete measures of the same size, one seeks a bijective assignment that minimizes some matching cost. For this task, exact algorithms are intractable for large problems, while approximate ones may lose the bijectivity of the assignment. We address this issue and the more general cases of non-uniform discrete measures with different total masses, where partial transport may be desirable. The core of our algorithm is a variant of the Quicksort algorithm that provides an efficient strategy to randomly explore many relevant and easy-to-compute couplings, by matching BSP trees in loglinear time. The couplings we obtain are as sparse as possible, in the sense that they provide bijections, injective partial matchings or sparse couplings depending on the nature of the matched measures. To improve the transport cost, we propose efficient strategies to merge k sparse couplings into a higher quality one. For k = 64, we obtain transport plans with typically less than 1% of relative error in a matter of seconds between hundreds of thousands of points in 3D on the CPU. We demonstrate how these high-quality approximations can drastically speed-up usual pipelines involving optimal transport, such as shape interpolation, intrinsic manifold sampling, color transfer, topological data analysis, rigid partial registration of point clouds and image stippling
Comprendre les tiers-lieux 'nourriciers' par la recherche participative
No full textInternational audienceCet article propose d'ouvrir le débat sur les définitions des tierslieux « nourriciers » à partir d'une recherche collaborative développée dans une ferme collective. Ce processus de recherche met en perspective la réflexivité et la participation des différentes parties prenantes (néo-paysans, salariés, chercheurs, éditeur) dans la co-construction du projet et son édition, afin d'affiner les savoirs sur cette initiative. Une expérience de science ouverte radicale qui nous permet, au croisement d'autres travaux, de proposer différentes configurations de tiers-lieux « nourriciers » en lien avec les transitions dont ils sont porteurs : les « tiers-lieux des transitions agri-alimentaires écologiques », qui se veulent orienté vers la production agricole et les consom'acteurs, et les « tiers-lieux de solidarité de transition alimentaire », qui se structurent dans l'accompagnement des processus de démocratisation.</div
Secure Federated Graph-Filtering for Recommender Systems
No full textRecommender systems often rely on graph-based filters, such as normalized item-item adjacency matrices and low-pass filters. While effective, the centralized computation of these components raises concerns about privacy, security, and the ethical use of user data. This work proposes two decentralized frameworks for securely computing these critical graph components without centralizing sensitive information. The first approach leverages lightweight Multi-Party Computation and distributed singular vector computations to privately compute key graph filters. The second extends this framework by incorporating low-rank approximations, enabling a trade-off between communication efficiency and predictive performance. Empirical evaluations on benchmark datasets demonstrate that the proposed methods achieve comparable accuracy to centralized state-of-the-art systems while ensuring data confidentiality and maintaining low communication costs. Our results highlight the potential for privacy-preserving decentralized architectures to bridge the gap between utility and user data protection in modern recommender systems
A tool for visual analysis and photo-realistic rendering of forest landscape model simulations
No full textInternational audienceSimulation outputs from Forest Landscape Models are complex, and tools for their visualanalysis and effective communication are often limited. In this paper, we present EcoViz, anovel, open-source visualization platform designed to complement existing forest models byproviding advanced 3D visualization capabilities. EcoViz facilitates the exploration ofsimulation results through two primary modes: symbolic rendering, designed for analyticaltasks, such as pattern recognition and model evaluation, and photorealistic rendering,leveraging physically-based rendering (Mitsuba 3) and a custom library of European 3D treemodels for communication purposes.The platform imports data in a typical model output format (e.g., spatially explicit individualtree or cohort data) and employs a temporally coherent sampling technique to visualizeindividual trees derived from cell-based density maps. Key features include: interactive side-by-side comparison of different simulation scenarios or time points, with synchronizednavigation (viewpoint, timeline, transects), a mini-map overview, timeline controls with linkedecological metric graphs, and transect analysis tools. Developed using C++/Qt/OpenGL,EcoViz is cross-platform.The practical application of EcoViz is demonstrated by visualizing simulations ofBerchtesgaden National Park under baseline and climate change scenarios exported fromthe iLand forest landscape model. This case study showcases the utility of EcoViz forcomparative scenario analysis across spatial scales and how it aids model evaluationthrough visual inspection. While symbolic views support detailed analysis, the photorealisticoutput offers a compelling tool for science communication with diverse audiences, includingscientific peers, forest managers, and the public
La poule en ville : représentations, pratiques et (re)qualification (XIXe-XXIe siècle)
No full texthttps://books.openedition.org/lisaa/238International audienc
Moving without cattle: migration patterns of Maasai from the Amboseli region (Southern Kenya)
No full textInternational audienceMaasai pastoralists face numerous constraints that compel them to diversify their livelihoods and, at times, migrate without their livestock. This paper focuses on such migrations undertaken by young Maasai men living in the Amboseli region of Kenya. The primary objective is to identify key migration patterns-namely, destinations, activities conducted, and the sequences of migration events. Fieldwork conducted in 2023 gathered information from 162 individuals who collectively undertook 445 migrations. Sequence and graph-based analyses were employed to explore these patterns and revealed key geographic hubs (e.g., Nairobi region and coastal areas), functional roles of locations (transit points versus places of residence), and typical migration trajectories. The analysis of graphs revealed three distinct forms: (i) one-way migrations, (ii) direct return trips, and (iii) circular migration patterns. A particularly noteworthy finding is the high frequency of return migrations among these individuals. Another important finding is the enduring attachment to the place of origin and pastoralism, maintained throughout migration trajectories and reinforced through remittances. This continued connection helps explain the high rate of return. Therefore, migration-far from alienating individuals from pastoralism-appears to serve as an instrumental means of sustaining it and eventually resuming it upon return.</div
« De quel droit ma jambe appartient-elle à Villemessant ? » Flaubert en lutte contre la circulation médiatique de ses biographèmes
No full textInternational audienc
Introduction générale: Éléments pour une histoire des rencontres entre syndicats et coopératives
No full textInternational audienc
From the 'Island Of Garbage' to Dark Maga : the Resurgence of a Masculinist Alt-Right in the 2024 Election?
No full textInternational audienceThis commentary explores the resurgence of masculinist movement and far-right ideologies in the 2024 US presidential election, focusing on Donald Trump's strategic use of these elements. Originating from digital spaces like 4chan, the Alt-Right movement emerges from phenomena such as #Gamergate while promoting the "red pill" concept: a metaphor for the awakening of online activists to perceived truths about societal structures, particularly gender and race. Trump's campaign leveraged this lexicon, appealing to young white male voters through associations with figures like Elon Musk and media personalities such as Joe Rogan. The emergence of "Dark Maga," a more extreme iteration of the MAGA movement, can be analyzed as a deeper radicalization of the movement, incorporating elements of the Neoreactionary (NrX) movement as well as building on the Alt-Right's ashes. This commentary highlights the broader implications of these developments, suggesting a potent political dynamic that is reshaping American politics
Post-pandemic changes in anxiety and depression symptom networks among socioeconomically disadvantaged young Adults: A repeated cross-sectional study
No full textInternational audienceBackgroundThe COVID-19 pandemic has significantly affected the mental health of young adults, particularly those facing socioeconomic hardship. Although psychological distress appears to be declining in the general population post-pandemic, vulnerable subgroups remain at elevated risk. Network analysis offers a transdiagnostic approach to understanding the dynamic interplay of depressive and anxiety symptoms over time.MethodsWe conducted a repeated cross-sectional study among socioeconomically disadvantaged young adults in 2020 (T1) and 2024–2025 (T2). Depressive and anxiety symptoms were assessed using the PHQ-9 and GAD-7. Symptom networks were estimated using Gaussian Graphical Models with EBICglasso. The Network Comparison Test (NCT) evaluated changes in network structure and symptom centrality. Clustering analysis was performed to explore the reorganization of symptom groupings over time.ResultsMean scores increased significantly between T1 (n = 960) and T2 (n = 380) for both depression (PHQ-9: 9.43 to 11.35, p < 0.001) and anxiety (GAD-7: 6.3 to 8.14, p < 0.001). Suicidal ideation nearly doubled (25.9 %–42.9 %, p < 0.001). Network analysis revealed stronger interconnections between depressive and anxiety symptoms at T2. Anxiety symptoms (particularly GAD3, GAD2, and GAD1) became more central, while suicidal ideation shifted from a depression-specific cluster to one integrating anxiety symptoms. Clustering analysis supported a progressive integration of depressive and anxiety domains.ConclusionOur findings suggest an evolving post-pandemic psychopathological network, with anxiety symptoms becoming increasingly central and closely linked to suicidal ideation. These results underscore the need for targeted interventions addressing both depression and anxiety, particularly among socioeconomically vulnerable young adults, to more effectively reduce suicide risk in this population