Fondazione Bruno Kessler

Memory-Efficient 3D Scene Understanding via Quantized Multi-View Features

Projecting high-dimensional representations derived with 2D foundation models onto 3D point clouds via multi-view aggregation has emerged as a powerful paradigm for 3D scene understanding. However, the conventional strategy of storing dense feature vectors - derived by 2D foundat...

Silicon Carbide Thin Films for MEMS: Correlating Vapor-Phase Deposition Parameters with Material Properties and Device Performance

Silicon carbide (SiC) thin films have emerged as a key material platform for micro-electromechanical systems (MEMS) designed to operate under extreme thermal, chemical, and mechanical conditions. While significant research has focused on individual SiC deposition techniques and m...

Construction and characterisation of the DarkSide-20k veto silicon photo-multiplier tiles

Abstract Silicon photo-multipliers (SiPMs) are state-of-the-art sensors capable of detecting a single photoelectron under cryogenic conditions, with potentially lower radioactivity than widely used photomultiplier tubes. The DarkSide-20k experiment, designed to perform direct dar...

Artificial Intelligence and Criminology

Machine learning is gaining momentum in criminology, and the unprecedented transformative power of generative artificial intelligence (AI) is poised to affect the discipline even more, perhaps in ways we currently struggle to imagine. In light of the growing role that AI plays in...

Quantum circuit-based learning models: bridging quantum computing and machine learning

Abstract Machine Learning (ML) has been widely applied across numerous domains due to its ability to automatically identify informative patterns from data for various tasks. The availability of large-scale data and advanced computational power enables the development of sophistic...

Multimodal large language models predict urban safety perception but encode non-neutral demographic priors

Understanding how people perceive urban environments is essential for inclusive planning, yet conventional surveys are costly and difficult to scale. We investigate whether Multimodal Large Language Models (MLLMs) can assess perceived urban safety from street-view imagery while a...

The LLM Wears Prada: Analysing Gender Bias and Stereotypes through Online Shopping Data

With the wide and cross-domain adoption of Large Language Models, it becomes crucial to assess the extent to which the statistical correlations in training data hide subtle and potentially troubling biases. Gender bias in LLMs has been widely investigated from the perspectives of...

Redox‐Confined Interfacial Polymerization Enables Mechanically Robust and Self‐Adhesive Nanofilms

ABSTRACT Robust polymer nanofilms are increasingly important for applications that require mechanically strong yet compliant membranes with nanoscale thickness and intimate interfacial contact. While biological membranes naturally combine flexibility, adaptability, and adhesion,...

Performance evaluation of a position-sensitive SiPM-based gamma camera for intraoperative imaging

Abstract Background The POSiCS camera is a handheld and wireless, small field-of-view gamma camera developed for use in radio-guided surgery (RGS), with sentinel lymph node biopsy (SLNB) as its benchmark application. This compact and lightweight detector (about $$350\,\textrm{g}$...

Silica-EmbeddedProteins Enable Higher Yield and AtomicIon Detection in Deep-UV Laser-Assisted Atom Probe Tomography

Abstract Determining the 3D structure of proteins is a long-standing challenge in the life sciences and is of paramount importance for understanding fundamental biological processes and advancing pharmaceutical development. Experimental techniques such as X-ray diffraction, nucle...