HAL Evry
Not a member yet
19237 research outputs found
Sort by
Eukaryote‐wide distribution of a family of longin domain‐containing GAP complexes for small GTPases
No full textInternational audienceArf and Rab family small GTPases and their regulators, GTPase‐activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), play a central role in membrane trafficking. In this study, we focused on a recently reported GAP for Arf (and potentially Rab) proteins, the CSW complex, a part of a small family of longin domain‐containing proteins that form complexes with GAP activity. This family also includes folliculin and GATOR1, which are GAPs for the Rag/Gtr GTPases. All three complexes are associated with lysosomes and play a role in nutrient signaling, the latter two being directly involved in the mTOR pathway. The role of CSW is not clear, but in addition to having GAP activity on Arf proteins in vitro, its mutation causes severe neurodegenerative diseases. Here we update the reported pan‐eukaryotic presence of folliculin and GATOR1, and demonstrate that CSW is also found throughout eukaryotes, though with sporadic distribution. We identify highly conserved motifs in all CSW subunits, some shared with the catalytic subunits of folliculin and GATOR1, that provide new potential avenues for experimental exploration. Remarkably, one such conserved sequence, the “GP” motif, is also found in structurally related longin proteins present in the archaeal ancestor of eukaryotes
: Curiosité intellectuelle, engagement académique, créativité sociologique
No full textInternational audienc
Recent Advancements in Humanoid Robot Heads: Mechanics, Perception, and Computational Systems
No full textInternational audienceThis paper presents a comprehensive review that provides an in-depth examination of humanoid heads, focusing on their mechanics, perception systems, computational frameworks, and human–robot interaction interfaces. The integration of these elements is crucial for developing advanced human–robot interfaces that enhance user interaction and experience. Key topics include the principles of context, functionality, and appearance that guide the design of humanoid heads. This review delves into the different aspects of human–robot interaction, emphasizing the role of artificial intelligence and large language models in improving these interactions. Technical challenges such as the uncanny valley phenomenon, facial expression synthesis, and multi-sensory integration are further explored. This paper identifies future research directions and underscores the importance of interdisciplinary collaboration in overcoming current limitations and advancing the field of humanoid head technology
Décisions d’investissement en main-d’œuvre dans les entreprises familiales : le comportement de sous-licenciement
No full textFNEGE 2International audienceThis study investigates labor investment inefficiency in family firms from a specific socioemotional wealth (SEW) perspective, particularly under-firing behavior. Family firms differ from their non-family counterparts and are more concerned about their reputation and emotional goals apart from financial objectives, leading them to make inadequate labor investment decisions. Based on a sample of French-listed firms in the SBF 120 index from 2005 to 2021, the results support the SEW perspective and show that family firms and their involvement in management lead to inefficient labor investment decisions, specifically through the under-firing problem. Additional analyses show that the over-investment problem is mitigated in family firms operating in a highly competitive environment. Moreover, inefficient labor investments are more pronounced in socially responsible family firms.Este estudio analiza la ineficiencia de la inversión en mano de obra en empresas familiares desde la perspectiva del capital socioemocional, centrándose en el comportamiento de subdespido. Motivadas por objetivos emocionales y de reputación, estas empresas toman decisiones laborales menos óptimas que las no familiares. Usando una muestra de empresas francesas del índice SBF 120 entre 2005 y 2021, los resultados confirman la influencia del capital socioemocional y muestran que la implicación familiar en la gestión acentúa estas ineficiencias. Sin embargo, en contextos de alta competencia, el problema del sobreempleo se reduce. Además, las empresas familiares con alta responsabilidad social presentan una mayor ineficiencia en sus decisiones de inversión en mano de obra.Cette étude examine l’inefficacité des investissements en main-d'œuvre dans les entreprises familiales sous l’angle spécifique de la richesse socio émotionnelle, en particulier le comportement de sous-licenciement. Les entreprises familiales, motivées par des objectifs émotionnels et réputationnels, prennent des décisions d’emploi moins optimales que les entreprises non familiales. À partir d’un échantillon d’entreprises françaises cotées à l’indice SBF 120 entre 2005 et 2021, les résultats confirment la perspective socio émotionnelle et montrent que l’actionnariat familial et l’implication de la famille dans la gestion de l’entreprise entraînent ces inefficiences notamment à travers le problème de sous-licenciement. Des analyses supplémentaires révèlent que la concurrence atténue cet effet. Par ailleurs, les investissements inefficaces en main-d'œuvre sont plus marqués dans les entreprises familiales socialement responsables
Data Beyond Control
No full textInternational audienceIn a context of increasing digitalization, controlling data flows constitutes a major challenge for organizations, particularly when it comes to sensitive data. This study examines the technical and legal challenges related to secure data flow management, based on the concrete case of the medical analysis laboratory MedLab and its facial recognition system implementation project. The study proposes an integrated approach combining Zero-Trust architecture, multi-factor authentication mechanisms, and advanced encryption solutions, while ensuring compliance with GDPR and other applicable regulations. Special attention is paid to subcontractor management and international data transfers. The results demonstrate that a holistic approach, combining innovative technical solutions and robust governance, can effectively meet security and compliance requirements. The study also emphasizes the importance of continuous adaptation to technological and regulatory developments, offering concrete recommendations for organizations wishing to strengthen control of their data flows
Data Fusion of Observability Signals to Detect Anomalies and Failures in Microservices
No full textInternational audienceMicroservices have become a reality, with large companies adopting this architecture due to its advantages, such as development agility, decoupling, scalability, resilience, and operational efficiency. However, this approach has drawbacks, such as the need for complex cloud infrastructure and communication between microservices through protocols. In organizations with hundreds of microservices, any failure in one of them can trigger cascading effects, resulting in more failures. Identifying failures in microservices is not a trivial task. The proposed work focuses on data fusion for detecting failures and anomalies in microservices, combining two pillars of observability: distributed tracing and metrics, with the help of OpenTelemetry. This fusion aims to provide faster and more efficient failure detection, as well as the identification of anomalies in metrics. This allows problems to be detected and addressed more quickly, improving the overall stability and reliability of the system
Shade-aware Routing For Sunburn Prevention
No full textInternational audienceWalking is a widely favored mode of travel, but pedestrians face significant exposure to ultraviolet radiation, which can cause skin discoloration and irritation. UV exposure can exceed healthy levels all year round, leading to short-term discomfort and long-term damage to the skin. The public is often advised to seek shade to avoid prolonged exposure to UV radiation. The practicality of this recommendation for pedestrians is not obvious as the shade coverage is highly variable in urban environments depending on the date and time. To test the feasibility of shade-aware routing, we model the approximate shade coverage with OpenStreetMap data and simulate pedestrian paths that minimize sun exposure. We quantify the opportunities to use shaded navigation in urban environments as it relates to avoidance of UV radiation. We show that it is useful to seek shade in urban environments even in situations where the shade coverage is the lowest. As an extreme case, we simulate shade on the day the UV index was the highest throughout 2024 in Paris, we show that it was 43% safer for pedestrians of skin type I to maximize shade as opposed to distance. Safe walking distance increased by 15%. Even on the day of minimal shade throughout the year, shade-aware routing was 34% safer compared to shortest-path routing, increasing safe walking distance by 4%
Gated Temporal Shifts with Depth-Efficient Channel Attention for Real-Time Hand-Gesture Interaction
No full textInternational audienceWe introduce a compact video-classification pipeline for real-time dynamic hand-gesture recognition in mixed-reality (MR) settings. The network marries a MobileNetV3 backbone with two purpose-built temporal components: (1) a Gated Discriminative Temporal Shift Module (G-DiTSM) that inserts first-order motion differences and learns channel-wise gates to fuse them adaptively, and (2) a lightweight Depth-Efficient Channel Attention (DepthECA) block that recalibrates spatial features on the fly. Operating on eight sparsely sampled frames per clip (Temporal Segment Network paradigm), the resulting model contains 2.65 M parameters and requires only 0.084 GFLOPs per inference. Evaluated on the RGB-only 20BN Jester benchmark (148k clips spanning 27 gesture classes) recorded from front-facing viewpoints. The system reaches 95.34% Top-1 and 99.80% Top-5 accuracy, surpassing recent 3D CNNs and transformer baselines while being an order of magnitude lighter. Ablations confirm that DepthECA and G-DiTSM provide complementary gains (+18.78% and +0.93% Top-1, respectively, over the MobileNetV3 baseline). Because all components are plug-and-play and introduce minimal overhead, the architecture is well suited to the tight latency and power budgets of standalone MR headsets, paving the way for natural grab, rotate, and command interactions using only on-board RGB cameras
The Hydrophilic Domain of HSulf Endosulfatases: An Intrinsically Disordered Region Governing Enzyme Functions and Therapeutic Potential
No full textInternational audienceAbstract Human extracellular endosulfatases HSulf-1 and HSulf-2 catalyze the selective 6-O-desulfation of heparan sulfate (HS), critically shaping the sulfation code that governs glycosaminoglycan (GAG)-mediated signaling. A unique feature of these enzymes is their hydrophilic domain (HD), an intrinsically disordered, non-conserved segment absent from all other human sulfatases. Despite lacking homology to known protein domains, the HD has emerged as a key regulatory module for substrate recognition, enzyme localization, and processive activity along HS chains. In this review, we dissect the structural and functional roles of the HD, with emphasis on its dynamic interaction with HS motifs and potential modulation of protein–GAG complexes. We also explore how its intrinsically disordered nature may confer conformational flexibility advantageous for navigating the complex landscape of extracellular glycans. Given the implication of HSulfs in diverse physiological and pathological contexts, including cancer, the HD presents a promising therapeutic target for the selective inhibition of endosulfatase activity. We discuss the challenges and perspectives in targeting intrinsically disordered regions (IDRs) in GAG-binding proteins, and highlight how the HD of HSulfs provides a paradigmatic example of non-canonical domains orchestrating fine-tuned GAG editing in the extracellular matrix