Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
Embeddings, topic models, LLM : un air de famille
No full textNational audienceWord embeddings, topic models, LLMs: a family affair This article presents a study on terms denoting family relationships (brother, aunt, etc.) in French using three approaches: word embeddings, topic modeling, and pre-trained language models. The first two types of representations are built from the French version of Wikipedia, while the third is derived through direct interaction with ChatGPT. The aim is to compare how these three methods represent such terms, in two main ways: by evaluating them against a structural definition of family relations (in terms of features such as gender, lineage, etc.), and by comparing the topics associated with each term. These methods reveal different modes of structuring family-related vocabulary, while also underscoring the continued necessity of corpus-based and controlled analyses to obtain reliable results
Towards Accurate Gait Instability Assessment: Identifying a Fixed Pelvis Point that Best Matches the Whole-Body Center of Mass
No full textInternational audienc
Pierrot lunaire à Paris (1921-1924): chronologie et dossier de presse des cinq représentations, dezede.org, https://dezede.org/dossiers/id/958/
No full texthttps://dezede.org/dossiers/id/958
Reconstruction of the regional dynamics of vegetation and human activities in the North-Western Alps over the last 7 millennia
No full textInternational audienceThe heterogeneity and complexity of mountain ecosystems can be attributed to a long history of interactions between a range of factors, including altitude and topography, as well as climate and human activities. Understanding the effects of climate and human activities on vegetation over time is crucial for assessing ecosystem changes and their long-term evolution. Holocene vegetation changes and human activities in the Alps have been well documented by pollen and sedimentary ancient DNA (sedaDNA) analyses. While these two proxies are relevant tools to characterise past anthropisation phases and their impact on vegetation at numerous sites across the Alps, they do not provide spatially explicit or quantitative estimates of changes in vegetation cover, nor of the proportion of land impacted by past human activities. The present study aims to reconstruct the spatial extent of the impact of human activities on the North-Western Alps landscape, based on sedimentary archives spanning the last 7 millennia. This has been achieved by using the Landscape Reconstruction Algorithm (LRA) sub-component REVEALS (Regional Estimates of VEgetation Abundance from Large Sites) on pollen data to reconstruct variations in plant abundance at a regional scale (radius of 50 km). Pollen-based vegetation modelling, such as the LRA, has been developed to correct for inter-taxonomic differences (pollen production, grain dispersal and deposition mechanisms). The REVEALS model was applied to eleven sites (nine lakes and two peat bogs) at altitudes ranging from 1,435m to 2,505m asl. The REVEALS outputs deliver regional dynamics of vegetation cover and human activities consistent with previous palynological and sedaDNA studies. They have highlighted already known anthropisation phases by revealing higher proportions of land-uses. They have also revealed new phases of anthropisation, that were unexpected based on pollen records alone. This enables us to reconsider the history and extent of the anthropisation of the different vegetation belts in the Alps. These regional-scale results, have been compared with an index of agro-pastoral activity intensity (based on pollen, spores and sedaDNA) and have therefore highlighted the variable spatial impact of human activities over time. Applying the LRA in this context enable us to identify the period when human influence, initially very local, became more spatially important, with a significant regional impact on the Alpine landscape from the beginning of the Middle-Ages around 1500 cal BP
Plus c’est petit, plus c’est grand. À propos des chefs-d’œuvre compagnonniques
No full textInternational audienc
Enriched environment requires the remodeling of hippocampal perineuronal nets to trigger memory improvement in Alzheimer’s mouse model
No full textInternational audienceAlzheimer’s disease (AD) is a major neurodegenerative disorder influenced by both genetic and environmental factors. Mentally stimulating activities are believed to reduce cognitive decline by establishing a cognitive reserve, although the underlying neurobiological mechanisms remain elusive. In this study, we investigate the role of parvalbumin-expressing inhibitory neurons (PV+) and their perineuronal nets (PNN) in the establishment of cognitive reserve using a mouse model of AD. We demonstrate that 10 days in an enriched environment (EE) significantly restores both spatial and social memory, which is accompanied by an increase in PV+ and PV+/PNN+ cells in the hippocampus. Notably, preventing PV/PNN remodeling in CA1 during EE abolishes the spatial memory improvements, whereas localized neuregulin-1 (NRG1) injections induce PV/PNN remodeling and restore memory performance. These findings indicate that hippocampal PV/PNN remodeling is a key contributor to the cognitive benefits of EE in AD, highlighting this neuronal population as a substrate for cognitive reserve
