Portail HAL UNIV-RENNES
Not a member yet
180610 research outputs found
Sort by
Identity and the Self. How Archaeology Creates Protohistoric Identities
No full textInternational audienc
Modular and adaptive implementation of Semantic Segmentation Models for Satellite Images and Open Source tools suitable for complex geographical contexts
No full textInternational audienceSemantic segmentation, the process of assigning a semantic label to each pixel in an image, is a critical computer vision task for extracting detailed information from remote sensing data. However, its application to complex geographical contexts, such as coastal wetlands, is often constrained by the need for highly specialized implementations, class imbalance, and limited accessibility for non-specialists. This paper introduces a novel, modular, and adaptive open-source framework for semantic segmentation tailored to satellite imagery. Designed for maximum flexibility, the framework supports both binary and multi-class segmentation tasks and incorporates specific training strategies to handle severe class imbalances inherent in ecological detection, such as salt marsh mapping. The implementation provides a fully configurable pipeline that bridges the gap between Geographic Information Systems (GIS) and Deep Learning (DL). It integrates QGIS for intuitive spatial preprocessing and grid generation with a Python-based training and prediction workflow, thereby democratizing access to advanced segmentation techniques. The framework is architecture-agnostic, allowing the seamless deployment and benchmarking of various state-of-the-art encoder-decoder models, which are effective at combining multi-scale contextual information with high spatial resolution. A key contribution is the integration of a multifaceted training methodology that includes hybrid loss functions with dynamic class weighting and spectral-consistent data augmentation to ensure robust model generalization from limited and imbalanced datasets. We demonstrate the framework's efficacy and scalability through two distinct case studies: a multi-class land cover classification on the Crozon Peninsula using Pléiades and a binary salt marsh detection in the Mont-Saint-Michel Bay Sentinel-2 imagery. The results show that accurate segmentation can be achieved with modest computational resources, promoting more sustainable and ethical AI applications in environmental monitoring. This work provides a critical tool for researchers and practitioners aiming to apply advanced DL segmentation to domain specific remote sensing challenges beyond conventional benchmarks
Novel genes arise from genomic deletions across the bacterial tree of life
No full textBacteria are hosts to enormous genic diversity. How new genes emerge, functionalize, and spread remain longstanding questions. Here, we explore a mechanism by which adaptive deletions fuse distant gene fragments. Unlike other gene birth mechanisms that begin with rare, neutral mutations, these "deletion-born fusions" reach high frequency by hitch-hiking on the deletion. The deletion-driven proliferation of the fusion prolongs the mutational supply within these genes before loss, providing additional opportunities for neofunctionalization. We document one such gene fixing and expressing in a long-term E. coli evolution experiment, and identify additional fusion events in the Mycobacterium tuberculosis-bovis split. Finally, we develop a scalable systematic screen to detect these genes in all 2.4 million public single-isolate genomes and identify deletion-born fusions across the bacterial tree of life. These findings challenge the notion that deletions are solely destructive and highlight their role as potential catalysts for evolutionary innovation
Comparison of antidepressant deprescribing strategies in individuals with clinically remitted depression: a systematic review and network meta-analysis
No full textInternational audienceBACKGROUND: Antidepressants are recommended for moderate-to-severe depression and anxiety, but concerns exist around overprescribing, long-term use, and paucity of evidence-based deprescribing strategies. We aimed to compare the effectiveness of different deprescribing approaches in individuals with clinically remitted depression or anxiety. METHODS: For this systematic review and network meta-analysis, we searched PubMed, PsycINFO, Web of Science, CENTRAL, CINAHL, and online trial registries from inception to April 6, 2025, for randomised controlled trials comparing abrupt discontinuation, fast tapering (≤4 weeks), slow tapering (>4 weeks), dose reduction (≤50% of the minimal effective dose), or continuation, with or without psychological support in adults with fully or partially remitted depressive or anxiety disorders on antidepressant treatment. The primary outcome was relapse rate by trial end. For each study, we extracted summary-level data on study characteristics, participant demographics, intervention details, and outcome measures. We did random-effects pairwise meta-analysis, and random-effects frequentist network meta-analysis to obtain relative risks (RRs) and standardised mean differences with 95% CIs. We assessed risk of bias using the Cochrane Risk-of-Bias-2 tool and the certainty of pooled estimates using the Confidence in Network Meta-Analysis approach. Individuals with lived experiences contributed to the interpretation of results. The study was registered with Open Science Framework, https://osf.io/9bsxz/. FINDINGS: Of 13 011 records, we included 76 trials comprising 17 379 participants. The mean age of individuals was 45·2 years (SD 15·2, IQR 34·9-55·5); 11 731 (67·5%) were female and 5648 (32·5%) male; the mean follow-up was 45·9 weeks (SD 29·7). Individuals were predominantly White (mean 87·9% [SD 8·1]). 60 (79%) of 76 studies investigated depression and 16 (21%) investigated anxiety. Separate analyses by diagnosis showed consistent baseline characteristics or treatment effects, although most comparisons were informed primarily by studies of participants with depression. After pooling data across conditions, the following strategies outperformed abrupt discontinuation for relapse prevention: continuation at standard dose plus psychological support (RR 0·40, 95% CI 0·26-0·61; number needed to treat [NNT] 4·3; moderate certainty), continuation at standard dose (0·51, 0·46-0·58; NNT 5·3; moderate certainty), slow tapering plus psychological support (0·52, 0·38-0·72; NNT 5·4; moderate certainty), and continuation at reduced dose (0·62, 0·42-0·92; NNT 6·8; low certainty). These strategies also outperformed fast tapering (point estimates ranging from 0·39 to 0·52; the first three supported by moderate certainty, the last by low certainty). Compared with abrupt discontinuation, the following strategies made no difference in relapse prevention: fast tapering plus psychological support (0·52, 0·27-1·01; low certainty), abrupt stopping plus psychological support (0·73, 0·30-1·78; very low certainty), and slow tapering alone (0·81, 0·56-1·18; low certainty). Moderate heterogeneity emerged (τ(2)=0·07), without evidence of inconsistency. Sensitivity, subgroup, and secondary outcomes provided consistent results. Tolerability was comparable across strategies. INTERPRETATION: In remitted depression, slow tapering plus psychological support is as effective as antidepressant continuation in preventing relapse and superior to abrupt or rapid discontinuation. In remitted anxiety, despite consistent population characteristics and effect estimates, limited evidence warrants cautious generalisation. Guidelines should promote individualised deprescribing with gradual tapering and structured psychological support. FUNDING: None
Effects of Pharmaceuticals and Personal Care Products on Insects
No full textInternational audiencePharmaceuticals and personal care products (PPCP) are omnipresent in our daily lives, and these emerging contaminants require increasing attention. PPCP and their metabolites are found in treated wastewater and in biosolids resulting from wastewater treatment, thus providing routes for dissemination to soils, freshwater, and groundwater, with subsequent consequences for insects. This review explores the effects of PPCP on the molecular, physiological, and biological (including behavior) responses of insects and consequences for their microbiota. As several PPCP can bioaccumulate in insects, in particular in aquatic ones, their roles in the transfer of the drugs to riparian and terrestrial food web are discussed, as well as community-level consequences. We propose future research directions to better appreciate the side effects and unintended effects of medication on insects. Attention should be given to the effects of PPCP on insects in environmentally realistic scenarios with relevant concentrations, in combination with other environmental pollutants, and with durations of exposure
Social and occupational factors are associated with musculoskeletal pain prevalence in the general population: a population-based cohort study
No full textInternational audienceObjective: Musculoskeletal pain (MP) is a leading cause of disability worldwide, affecting individual well-being and public health. However, in the literature, the prevalence of MP varies considerably because of methodological inconsistencies, selection biases, and differences in case definitions. This study aimed to estimate the population-based prevalence of MP in France and identify key demographic, socioeconomic, and occupational factors associated with MP.Methods: This cross-sectional study used baseline data for the CONSTANCES cohort study, a large, population-based epidemiological study with participants representative of the French adult population (18-69 years old). Inverse probability weighting was used to correct for selection bias and to improve the generalizability of prevalence estimates. MP was were assessed with the Nordic Musculoskeletal Questionnaire, with significant pain defined as lasting >30 days in the past 12 months. Multivariate logistic regression models were used to identify factors associated with low back pain, estimating odds ratios (ORs) and 95% confidence intervals (CIs).Results: Among 193,436 participants, 46.2% reported pain in at least one anatomical site. The most affected areas were the low back (26.6% adjusted prevalence), shoulder (21.4%), neck (19.0%), and knee (19.1%). Odds of low back pain was associated with female sex (OR 1.39 [95% CI 1.32-1.47]), older age, obesity, depression (1.71 [1.62-1.80]), and comorbidity burden (1.20 [1.15-1.25]). Odds of low back pain was associated with moderate or high occupational physical activity (OR 1.33 [1.20-1.50] and 1.69 [1.48-1.93]) but was inversely associated with very active leisure-time physical activity (0.82 [0.70-0.96]). Education level but not household income was a significant socioeconomic factor associated with MP.Conclusion: MP impose a substantial burden on the French population, particularly among individuals with physically demanding jobs and low education levels. These findings highlight the paradox of physical activity associated with MP
Online measurement of differential drain to source voltage for SiC MOSFETS power module
No full textInternational audienceThe on-state resistance RDSON is a key aging indicator of silicon carbide MOSFETs power module. It can provide information on both chip and packaging degradation allowing more reliable power electronic converter. This on-state resistance can be determined using both the on-state current and voltage of the power semiconductor. This paper focuses on an on-state voltage measurement circuit for SiC MOSFET power module. In the literature, power module bondwires resistance cannot be measured due to high switching voltage oscillations. An on-state voltage measurement circuit that can withstand such oscillations is proposed. The proposed circuit is characterized using the ISO 5725 standard over a wide temperature range. Finally, the circuit is successfully tested on a double pulse bench for various module base plate temperatures and compared to data obtained with a static curve tracer Keysight B1505
Representability of the direct sum of uniform q-matroids
No full textThere are many similarities between the theories of matroids and -matroids. However, when dealing with the direct sum of -matroids many differences arise. Most notably, it has recently been shown that the direct sum of representable -matroids is not necessarily representable. In this work, we focus on the direct sum of uniform -matroids. Using algebraic and geometric tools, together with the notion of cyclic flats of -matroids, we show that this is always representable, by providing a representation over a sufficiently large field
Search for planetary-mass ultra-compact binaries using data from the first part of the LIGO--Virgo--KAGRA fourth observing run
No full textInternational audienceWe present a search for gravitational waves from inspiraling, planetary-mass ultra-compact binaries using data from the first part of the fourth observing run of LIGO, Virgo and KAGRA. Finding no evidence of such systems, we determine the maximum distance reach for such objects and their merger rate densities, independently of how they could have formed. Then, we identify classes of primordial black-hole mass distributions for which these rate limits can be translated into relevant constraints on the mass distribution of primordial black holes, assuming that they compose all of dark matter, in the mass range . Our constraints are consistent with existing microlensing results in the planetary-mass range, and provide a complementary probe to sub-solar mass objects
Biopouvoir en culture de l’écran
No full textInternational audienceLa notion de biopouvoir défi nie par Foucault comme le contrôle de la vie par une forme de pouvoir souvent (mais pas seulement) incarnée par l’État, a fait preuve de sa pertinence lors de la pandémie de COVID-19. Cet ouvrage analyse plus particulièrement les formes multiples du biopouvoir, les résistances et les adhésions qu’il peut susciter. Il tente de comprendre leur mise en œuvre dans la culture de l’écran telle que vécue par nos sociétés contemporaines. Littératures anglophones et productions télévisuelles, de Margaret Atwood à Don DeLillo, de Desperate Housewives à Westworld, côtoient ainsi des analyses des systèmes de gestion d’épidémie, de reconnaissance faciale, et des figures tutélaires de la fiction comme le monstre ou le zombie. Le livre permet de réfléchir aux formes du biopouvoir, dans ses manifestations corporelles, sociales, numériques, et culturelles, à la fois sous l’angle de la discipline et de la biopolitique