Université de technologie de Troyes open archive
Not a member yet
10722 research outputs found
Sort by
Fluorescent nanoparticles for security tags against the looting of archaeological artefacts
No full textInternational audienceSeveral security solutions exist to reduce the risk of looting and theft of cultural artefacts, including systematic tagging to deter illicit trade. This study explores the integration of fluorescent nanoparticles with unique optical properties into ready-to-use conservation varnishes for archaeologists. Various nanomaterials (carbon dots, chalcogenides, and perovskite quantum dots) were incorporated into conservation resins to develop diverse tagging solutions compatible with archaeological artefacts. Their stability was assessed through accelerated aging by comparing their initial fluorescence brightness and its reduction after aging. Paraloid-based tags provided superior protection to alternative resins, while perovskite quantum dots exhibited instability under harsh conditions compared to other nanomaterials. Potential interferences with composition analysis were also evaluated by determining which tags could be detected using ED-XRF and WD-XRF. This study complements existing security tagging methods and may contribute to the broader adoption and improvement of traceable tagging solutions
Differential modulation of endothelial cell functionality by LRP1 expression in fibroblasts and cancer-associated fibroblasts via paracrine signals and matrix remodeling
No full textInternational audienceLRP1 is a multifunctional endocytosis receptor involved in the regulation of cancer cell aggressiveness, fibroblast phenotype and angiogenesis. In breast cancer microenvironment, cancer-associated fibroblasts (CAFs) play a crucial role in matrix remodeling and tumor niche composition. LRP1 expression was described in fibroblasts and CAFs but remains poorly understood regarding its impact on endothelial cell behavior and angiocrine signaling. We analyzed the angio-modulatory effect of LRP1 expression in murine embryonic fibroblasts (MEFs) and breast cancer-educated CAF2 cells. We employed conditioned media and fibroblast-derived matrices to model fibroblastic cells angiogenic effects on human umbilical vein endothelial cells (HUVEC). Neither the extracellular matrix assembled by MEFs knock-out for LRP1 (PEA-13) nor their secretome modify the migration of HUVEC as compared to wild-type. Conversely, LRP1-deficient CAF2 secretome and matrices stimulate endothelial cell migration. Using spheroids, we demonstrate that PEA-13 secretome does not affect HUVEC angio-invasion. By contrast, CAF2 secretome invalidated for LRP1 stimulates endothelial sprouting as compared to controls. In addition, it specifically stabilized peripheral VE-cadherin-mediated endothelial cell junctions. A global proteomic analysis revealed that LRP1 expression in CAFs orchestrates a specific mobilization of secreted matricial components, surface receptors and membrane-associated proteins at the endothelial cell surface, thereby illustrating the deep influence exerted by LRP1 in angiogenic signals emitted by activated fibroblasts
Sustainable Routing for Recyclable Waste Collection in Low- and Middle-Income Countries
No full textInternational audienc
Driving Control in Autonomous Vehicles: An AI-Empowered Safety-Centric Approach
No full textInternational audienceEnsuring safety in autonomous driving requires robust systems capable of detecting and mitigating dangerous driving behaviors in real time. This paper proposes an AI-powered approach that integrates deep learning models, specifically Convolutional Neural Networks (CNNs), with sensor fusion from vehicle-embedded sensors and onboard cameras. The system processes multimodal data to identify hazardous autonomous driving patterns and applies a Random Forest classifier to distinguish between normal and unsafe behaviors. By leveraging advanced data fusion and machine learning techniques, the proposed method enhances the reliability of autonomous driving safety mechanisms. Experimental results demonstrate the effectiveness of the approach in accurately classifying driving behaviors, offering a promising solution for proactive risk mitigation and accident prevention in autonomous vehicles.</div
Towards Sustainable Knowledge Management: The Emergence of Non-Pre-Competitive Knowledge Ecosystems in a Regulated Safety Context
No full textInternational audienceThis study presents a qualitative case study exploring how inter-organizational knowledge management is structured within a regulated safety context, where the core objective is the long-term preservation of critical knowledge. Within such a context, organizations collaborate to collectively create, share, and sustain knowledge over time. Our analysis shows that the most effective response to this challenge lies in the reconfiguration of inter-organizational architecture into a non-pre-competitive knowledge ecosystem. This type of ecosystem, still underexplored in the literature, is characterized by the absence of commercial or anticipatory competitive goals. It is structured around the production, preservation, and emergence of interconnected knowledge systems, supported by organizational innovation involving new practices, structures, and coordination mechanisms.Adopting an organizational perspective that bridges insights from knowledge management and organizational innovation, we analyze the emergence of such an ecosystem through the case of Kronos, a network of organizations operating in a highly regulated environment. The case highlights how relationship formalization, institutional regulation, and the presence of a focal actor contribute to a collective goal of knowledge transmission, sustainability, and systematization. This paper offers a conceptualization of non-pre-competitive knowledge ecosystems, emphasizing their distinctiveness from traditional pre-competitive frameworks and underlining the structuring role of organizational innovation in sustainable knowledge management
A Hybrid Autoencoder-Transformer Model for Detection of Attacks on Low Latency Services
No full textInternational audienceThis paper addresses the problem of detection of attacks in computer networks. More precisely, we consider attacks on emerging low-latency services, which typically require a specific traffic management system. We present a simple yet very efficient hybrid method that takes advantage of both autoencoder and transformer models. The original method is compared with the current state-of-the-art on a large real-life dataset of network traffic to show the relevance of the proposed approach, especially for low false-positive rates. A quick ablation analysis shows that the efficiency of the method relies on the combinaison of the two methods jointly used in our hybrid model
Evaluation methodology and application of Scope 3 emissions at the city and region scale in South Korea
No full textInternational audienc
Evaluating waste management performance in South Korea: insights over 27 years
No full textInternational audienc
Defining the teaching profiles of academic staff across a European universities alliance: Lessons learned after the pandemic and the way into the future
No full textInternational audienceThis study examines the teaching profiles of academic staff across a European universities alliance, explores the lessons learned after the pandemic, and outlines the path forward. Improvements in higher education systems and practices that embrace digital transition and equip academics with the necessary skills to facilitate quality learning are necessary in today's rapidly changing societies. As the COVID-19 pandemic swept across the world, educational institutions were forced to adapt their teaching practices in a state of urgency. The European University of Technology (EUt+) was not immune to the pandemic's impact. This study focused on how the eight universities of EUt+ responded to the pandemic, adapting their teaching and assessment practices. The study aimed at drawing the teaching profiles of the staff, exploring the different teaching modes before, during, and after the pandemic, and ways in which academics can exchange knowledge and value experiences related to the teaching process. The study followed the conventions of exploratory research, employing a mixedmethods approach; the data were obtained through an electronic questionnaire sent to all the members of the staff across all eight universities of EUt+ and semi-structured focus-group sessions. Findings showed that before COVID-19 the majority of the members of academic staff delivered their classes through lectures, seminars, and tutorial interaction, while traditional types of assessment such as closed book exams, project work, group work, or practical work were frequently used. Furthermore, more conventional technology tools were integrated into their teaching practice rather than new and emerging technologies. Nevertheless, the pandemic brought about several changes both in the teaching and assessment methods, shifting attention to tasks that required more use of critical-thinking skills and the challenge of limiting plagiarism.</div