Archivio della ricerca - Fondazione Bruno Kessler
Not a member yet
    21227 research outputs found

    High-Resolution Open-Vocabulary Object 6D Pose Estimation

    No full text
    : The generalisation to unseen objects in the 6D pose estimation task is very challenging. While Vision-Language Models (VLMs) enable using natural language descriptions to support 6D pose estimation of unseen objects, these solutions underperform compared to model-based methods. In this work we present Horyon, an open-vocabulary VLM-based architecture that addresses relative pose estimation between two scenes of an unseen object, described by a textual prompt only. We use the textual prompt to identify the unseen object in the scenes and then obtain high-resolution multi-scale features. These features are used to extract cross-scene matches for registration. We evaluate our model on a benchmark with a large variety of unseen objects across four datasets, namely REAL275, Toyota-Light, Linemod, and YCB-Video. Our method achieves state-of-the-art performance on all datasets, outperforming by 12.6 in Average Recall the previous best-performing approach

    Performance of DC resistive read-out silicon sensors for future 4D tracking

    No full text
    Abstract In recent years, the innovative concept of low internal gain avalanche detector (LGAD) coupled with the resistive read-out (RSD) has significantly changed the performance of silicon detectors. By increasing the signal-to-noise ratio by a factor of around 20, the LGAD mechanism led to unprecedented time resolution, on the order of 30 ps for a 50 μm active sensor thickness. The resistive read-out allows instead to share the collected charge between several electrodes, achieving an excellent spatial resolution, order of 5% of the pixel pitch. These two concepts combined enable 4D tracking with silicon sensors. This contribution presents the latest silicon sensor design based on these two concepts: a thin LGAD with a DC-coupled resistive read-out, the DC-coupled Resistive Silicon Detector (DC-RSD). This design aims at achieving controlled signal sharing through the implementation of isolating trenches (TI technology) amongst the electrodes that define the pixel. Various structures are implemented in the wafer layout, manufactured at FBK as part of the 4DSHARE project. A subset of sensors was characterised with a laser Transient Current Technique (TCT) system, equipped with a 5-μm spot, to estimate the ultimate spatial resolution as a function of the pitch. Measurements at the TCT are also performed to evaluate how different values of the resistive layer impact the signal sharing. The spatial resolution of sensors tested at DESY with an electron beam will also be presented. This study will provide significant feedback on the capabilities and reliability of the DC-RSD design

    Timing resolution of thin LGAD sensors in high radiation environments

    No full text
    Abstract The EXFLU batches of thin LGAD sensors with substrates of thickness between 20 and 45 μm were tested for timing performance at the DESY Test Beam Facility. The sensor designs are optimised for excellent timing resolution, and are manufactured by Fondazione Bruno Kessler. The timing performance is reported for both non-irradiated LGAD sensors and a set of 30 μm sensors irradiated up to 2.5 × 10 15  n 1 MeV eq. cm -2 . In addition, a new approach to the analysis of thin sensors is considered using differential programming to optimise a signal probability model. The results using the conventional method and this approach using machine learning are presented

    Chatbot Confessions: Large-Scale Analysis of Private Data Disclosure in Shared AI Chatbot Conversations

    No full text
    The proliferation of AI conversation platforms has introduced unprecedented privacy risks through user-shared conversations. This paper presents a comprehensive analysis of privacy vulnerabilities in shared conversations across three major LLM platforms: ChatGPT, Microsoft Copilot, and Google Gemini. We collected and analyzed 100 342 conversations using an automated LLM-based privacy detection pipeline enhanced with a defined risk scoring system and the LINDDUN threat modeling framework. Our analysis identifies 8 131 conversations (8%) to incur privacy risks deriving from the disclosure of private and sensitive data including user identifiers (49%) and user location data (40%), yet in some cases also financial (4%), health (3%) and authentication data such as access tokens (3%). Through systematic analysis of conversation length and temporal disclosure patterns, we demonstrate that extended conversations exhibit higher privacy risk rates compared to brief interactions. Notably, 60% of private data disclosures in longer conversations occur in the final quartile of these conversations, which may indicate that users progressively lose privacy awareness as interactions deepen. Our findings have immediate implications for platform designers and policymakers, highlighting the need for proactive interventions including real-time privacy warnings, preshare scanning, and clearer education about the permanence and discoverability of shared conversation links

    Lutter hors des partis. La gauche ­révolutionnaire ­italienne des ­années 1960-1980

    No full text
    L'ouvrage rassemble les témoignages inédits de cinq hommes et cinq femmes impliqué·es dans des organisations de la gauche extra-parlementaire italienne des années 1960-1980, parmi lesquels ceux de Toni Negri et de Sergio Segio. À partir de ces récits biographiques, l’ ouvrage fait ressortir les facteurs favorisant l’ engagement illégal ou clandestin dans un contexte de forte tension et de violence politique, y compris de la part de l’ État. Ce livre n’ a pour vocation ni de condamner ni de faire l’ apologie des positions ou des actions entreprises par les groupes considérés et les activistes interviewés. L’ objectif est ici de reconstruire un sens à partir de la parole des acteurs dont on cherche à saisir les ressorts de la trajectoire. Offrant une lecture vivante, l’ ouvrage s’ adresse au grand public. En lui faisant entendre des voix restées jusqu’ ici peu audibles, il lui livre non seulement des données utiles à une meilleure compréhension partagée de la période historique évoquée mais aussi des éléments qui éclairent, à bien des égards, l’ époque actuelle

    1,106

    full texts

    21,227

    metadata records
    Updated in last 30 days.
    Archivio della ricerca - Fondazione Bruno Kessler
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇