513225 research outputs found
Sort by
Phascolarctobacterium faecium reduces food intake via pyy signaling, contributing to the mitigation of body weight gain in diet-induced obese mice
No full textExcess energy intake contributes to adiposity in obesity. We investigated whether the human intestinal bacterium Phascolarctobacterium faecium could prevent obesity via enteroendocrine pathways in a mouse model of diet induced obesity (DIO). Daily administration of P. faecium (2 × 109 cells/mouse) reduced food intake through the early overproduction of the satiety hormone peptide YY (PYY) compared to untreated DIO mice. Moreover, P. faecium increased the intestinal levels of branched-chain amino acids, which, in turn, stimulated PYY secretion in neuroendocrine cell cultures and also modified gut microbiota composition. A pair-feeding study demonstrated that the anorexigenic effect of P. faecium contributes to its effects in attenuating body weight gain in DIO mice, but that other mechanisms are also involved in its metabolic benefits. Specifically, P. faecium accelerated gut transit and serum lipid clearance, thereby limiting adiposity independently of food intake. This study identifies the mode of action of a human intestinal bacterium recently linked to obesity protection, providing valuable insights into host-microbe interactions governing body weight
Mercury treatment in late medieval European leprosaria? New data from human dental calculus
No full textThis study investigates the potential use of mercury-based treatments for leprosy during the late medieval period through the analysis of mercury concentrations in human dental calculus. Calculus samples were obtained from individuals buried at St Leonard's (Peterborough, England) and St Thomas's (Aizier, France) leprosaria, with comparative samples collected from non-leprosaria burial sites where individuals showed no osteological evidence of leprosy. Mercury levels were measured in soil (N = 45) from the burial grounds using atomic absorption spectrometry and in dental calculus (N = 76) using cold vapour fluorescence adsorption spectroscopy. The findings reveal significant variability in mercury concentrations within and between sites, with the dental calculus values ranging from 0.11 to 9.7 mg kg−1. Notably, individuals from leprosaria cemeteries exhibited significantly higher mercury concentrations in their calculus than those from non-leprosaria cemeteries (t(70) = 3.22, p = 0.002). Although mercury levels in the burial soils associated with the individuals sampled in leprosaria cemeteries were elevated compared to the natural geochemical background, the analysis suggests that contamination is probably due to mercury accumulated in soft tissues during life, which could substantially increase the surrounding soil's mercury content during decomposition. Thus, contamination from the surrounding soil was ruled out as the source of mercury in the dental calculus. Integrated with historical and archaeological evidence, these findings may suggest the intentional use of mercury-based treatments for leprosy. This study highlights the role of bioarchaeology in uncovering medieval therapeutic practices and provides the first evidence of mercury linked to medical treatment detected in dental calculus
Exploring the endonuclease MUS81 as a new factor impacting on cancer chemosensitivity
No full textThe MUS81 complex is a crucial factor for genome stability maintenance and has been proposed as cancer predisposition gene 1,2. The complex is regulated by Casein Kinase 2 (CK2) through phosphorylation at serine 87 (S87), which is essential for resolving branched intermediates during mitosis 3. Our previous findings show that the phosphomimetic S87D-MUS81 mutant confers PARP inhibitor (PARPi) resistance in BRCA2-deficient backgrounds 4, possibly because replication-related double-strand breaks (DSBs) can be completely resolved through alternative end-joining (Alt-EJ) 5. Thus, deregulated function of MUS81 emerged as a potential modifier of the response to PARPi when BRCA2 is mutated or absent. In this study, we investigate why the S87D-MUS81 mutant aberrantly targets DNA replication forks. We found that S87D-MUS81 is hyper-recruited at ongoing replication forks and that the characteristic phenotype of S87D-MUS81 mutant cells is reversed by depletion of the protein scaffold SLX4 by siRNA transfection. This indicates that the unscheduled cleavage function conferred to MUS81 by the S87D mutation and excessive recruitment at DNA replication forks depend on SLX4. We confirm these findings also using a set of mutations that abrogate the binding of MUS81 to SLX4. These findings could lead to a better understanding of the relationship between MUS81 status, PARPi chemosensitivity and a possible MUS81 exploitation in target therapy
Efficient deep learning for computer vision
No full textThe advancement of modern computer vision, while driven by the remarkable performance of Deep Neural Networks (DNNs), is increasingly confronted by the practical limitations imposed by their own escalating complexity. A prevailing trend toward ever-larger and more computationally demanding models has created a significant barrier to the deployment of sophisticated artificial intelligence on resource-constrained platforms, such as mobile devices, wearables, and autonomous drones. This thesis directly confronts this critical challenge by investigating and advancing techniques for efficient deep learning, with a central focus on two powerful and aggressive model compression strategies: Binary Neural Networks (BNNs) \cite{courbariaux2016binarized} and neural network pruning. The core hypothesis guiding this work is that extreme model compression can be achieved while maintaining high performance across a diverse set of real-world computer vision problems, thereby enabling the deployment of advanced AI in practical, resource-limited scenarios.
Methodologically, this research first addresses the foundational problems that have impeded the widespread adoption of these efficient models. This includes the development of CycleBNN \cite{cyclebnn}, a novel cyclic precision training methodology designed to significantly reduce the substantial computational overhead associated with training BNNs. In a complementary investigation into sparsity, this work also introduces Distilled Gradual Pruning with Pruned Fine-tuning (DG2PF) \cite{gradual_pruning}, a comprehensive algorithm that synergizes network pruning with knowledge distillation to achieve high levels of model compression with minimal degradation in accuracy. Having established these core techniques, the thesis proceeds to demonstrate their practical viability through their application to a series of challenging, latency-sensitive tasks. Highly efficient BNN-based solutions are presented for hand gesture recognition (BNNAction-Net)\cite{bnnaction}, deepfake detection (Faster Than Lies) \cite{lanzino2024faster}, and cross-view geolocalization for UAVs (BiCrossNet) \cite{bicrossnet}. In each case, the proposed models achieve performance competitive with their full-precision counterparts while offering reductions in computational complexity and memory footprint that span orders of magnitude.
Furthermore, this research extends the study of deepfake detection beyond the paradigm of static classification by reframing it as a continual learning problem. Through a novel chronological evaluation framework that simulates the real-world evolution of generative technologies, a fundamental limitation in the generalization capabilities of current detectors is identified. This critical finding leads to the proposal of the Non-Universal Deepfake Distribution Hypothesis\cite{fontana2025revisiting}, which posits that each deepfake generator imprints a unique, non-transferable signature, thereby underscoring the absolute necessity of continuous model adaptation for any robust, long-term detection strategy.
The collective results of this thesis empirically validate that BNNs and neural network pruning are not merely theoretical concepts but are powerful, practical tools for developing high-performance, resource-efficient computer vision systems. By systematically bridging the gap between state-of-the-art accuracy and real-world deployability, this work contributes to a more accessible, sustainable, and scalable future for the field of artificial intelligence
La specializzazione metallurgica nelle comunità terramaricole: un approccio traceologico e di archeologia sperimentale
No full textIl progetto di ricerca indaga l’organizzazione della produzione metallurgica nelle comunità della cultura delle Terramare, con particolare attenzione al tema della specializzazione artigianale. Lo studio si basa su una selezione di manufatti in bronzo provenienti dai siti di Pragatto (BO), Santa Rosa di Poviglio (RE) e San Michele di Valestra (RE). L’obiettivo principale è comprendere l’organizzazione dell’attività metallurgica all’interno delle comunità terramaricole e valutare l’esistenza di differenti livelli di specializzazione nella produzione dei manufatti metallici.
Attraverso l’analisi delle tracce tecnologiche e funzionali presenti sui reperti, il progetto mira a ricostruire i gesti tecnici e le competenze necessarie alla fabbricazione delle diverse classi di oggetti in bronzo. La distribuzione e la morfologia delle tracce consentono di interpretare il grado di abilità dell’artigiano e di ipotizzare una possibile articolazione interna del sistema produttivo, basata sull’esperienza, sulle capacità tecniche e sulle fasi di lavorazione. L’ampia diffusione del metallo nelle comunità terramaricole, attestata sia da manufatti di uso comune sia da produzioni di maggiore pregio, suggerisce l’esistenza di forme di organizzazione produttiva complesse e di artigiani dedicati in modo continuativo alla lavorazione dei metalli.
Il confronto tra differenti classi di materiali e tra contesti diversi permette di evidenziare scelte tecnologiche e culturali condivise o differenziate, contribuendo alla definizione dei modelli di produzione metallurgica intra e inter-site. Il progetto intende così offrire nuovi elementi di riflessione sul fenomeno della specializzazione metallurgica nelle società protostoriche.This research project investigates the organization of metallurgical production in Terramare culture communities, with a particular focus on artisanal specialization. The study is based on a selection of bronze artifacts from the sites of Pragatto (BO), Santa Rosa di Poviglio (RE), and San Michele di Valestra (RE). The primary objective is to understand the organization of metallurgical activity within Terramare communities and assess the existence of different levels of specialization in the production of metal artifacts.
Through the analysis of technological and functional traces present on the artifacts, the project aims to reconstruct the technical skills and abilities required to produce the various classes of bronze objects. The distribution and morphology of the traces allow us to interpret the level of skill of the artisan and to hypothesize a possible internal structure of the production system, based on experience, technical ability, and processing phases. The widespread use of metal in Terramare communities, evidenced by both everyday artifacts and more valuable productions, suggests the existence of complex forms of production organization and artisans continuously dedicated to metalworking.
Comparisons between different classes of materials and across diverse contexts highlight shared or differentiated technological and cultural choices, contributing to the definition of intra- and inter-site metallurgical production models. The project thus aims to offer new insights into the phenomenon of metallurgical specialization in protohistoric societies
The European ‘Regulatory Turn’ towards Big Tech Platforms through the Lens of Digital Constitutionalism
No full textThis paper will draw connections between the development of internet regu lation, the theoretical foundations of digital constitutionalism, and the recent ‘regulatory turn’ endeavoured by the European Union, as evidenced by the Digital Markets Act and the Digital Services Act, both approved in 2022. The research will examine how internet and platform regulation have emerged and have been shaped over the past decades, and what the similarities and differences between US and EU approaches to the issue are. Special emphasis will be placed on the freedom of expression in the digital environment, which serves as a key test for the extent to which rights and freedoms are being com promised by platform dynamics. Finally, the paper will examine Europe’s new constitutional approach to internet and platform regulation, which marks a step forward in a complex strategy to manage the digital environment while preserving European democratic values and fundamental rights
Unitary and homotopy equivalences: classification of low-dimensional topological phases of quantum matter
No full textTopological insulators have attracted significant attention across physics and mathematics due to their technological potential and their rich geometric and algebraic structure. Their hallmark property is the bulk–boundary correspondence, whereby non-trivial bulk topology enforces the existence of metallic edge states. The mathematical classification of such phases of matter, culminating in the periodic table of topological insulators and superconductors, associates topological invariants to different symmetry classes defined by time-reversal, particle–hole, and chiral symmetries. However, this framework involves several approximations: it replaces projection-valued maps (PVMs) with vector bundles, collapses the torus to a sphere -- thus overlooking weak, lower-dimensional invariants --, and employs stable equivalence notions from K-theory that do not always capture the full topological content.
This thesis develops a direct homotopy-theoretic approach to the classification of symmetric PVMs, aimed at overcoming these limitations. In particular, it investigates the interplay between unitary equivalence and homotopy, identifies weak invariants absent from the Kitaev table, and establishes a general classification scheme applicable to arbitrary periodic models. The analysis focuses on low-dimensional settings (0, 1, and 2-dimensional systems), which provide a tractable yet non-trivial testing ground. Within this framework, we examine models with a single symmetry present, corresponding to the Altland–Zirnbauer classes A, AI, AII, AIII, C, and D. The results clarify in particular the obstructions to constructing symmetric Wannier bases, refine the understanding of strong versus weak topological invariants, and detail the relation between topological phases of matter and the dimerization choice in discrete models
Preserving the Relationship Between Art and Architecture. A Mixed Reality Application for the ‘Gallery of Rare Paintings’ in the Church of San Pietro in Valle (Fano, Italy)
No full textExtended reality is increasingly being used to disseminate cultural content, enabling innovative storytelling and experiences. Especially, Mixed Reality (MR) allows for the seamless integration of between virtual and real environments, transforming cultural sites into immersive multidimensional palimpsests of digital content. Building on these premises, the paper presents a MR experience of the Baroque church of San Pietro in Valle in Fano (Marche, Italy). The research intends to digitally re-establish the link between the art and architecture of the church, which was lost due to various historical events that deprived the building of its ‘gallery of rare paintings’. To achieve this goal, an on-site digital enhancement strategy is designed, implementing a MR environment that ensure the visitors to experience the church’s original art collection virtually relocated on site. Using a forefront MR headset, users can interact with the repositioned artworks, including pieces by Guido Reni, Simone Cantarini, and Matteo Loves and enrich their cultural encounters with interactive, immersive experiences. The result is a highly immersive cultural interaction that recreates value-based relationships, ensuring an experience of extended accessibility focused on heritage space and memory
Autophagy and mitophagy in dermatological disease: a comprehensive review from molecular pathways to therapeutic frontiers
No full textAutophagy - the cell's built-in recycling and quality-control programme - touches every layer of cutaneous biology. In keratinocytes it sculpts the cornified envelope; in melanocytes it balances pigment synthesis and oxidative stress; in immune and appendageal cells it fine-tunes defence, repair and hair-follicle cycling. When this choreography falters, skin disorders emerge. This review journeys from basic mechanisms (ULK1 signalling, Beclin-1/VPS34 nucleation, LC3B lipidation, selective mitophagy) to their fingerprints in health and disease. We dissect how autophagy malfunctions drive psoriasis hyper-proliferation, atopic-dermatitis barrier leakiness, vitiligo depigmentation and the metabolic rewiring of melanoma. Non-melanoma cancers, infectious dermatoses, wound repair, ageing and photo-damage are mapped onto the same autophagic atlas. Therapeutically, the pathway is a double-edged sword. mTOR or caloric-restriction mimetics jump-start a protective flux; chloroquine derivatives and ULK1 blockers clip tumour survival circuits; cannabinoids, photodynamic therapy and immune-checkpoint combinations exploit context-specific toggling between induction and brake. Emerging biomarkers (LC3B-II, p62, AMBRA1) promise patient-stratified interventions. By weaving together molecular detail, pre-clinical insight and clinical translation, we show why autophagy is no longer a backstage process but a star player in dermatology - and how targeting its switches could reshape future treatment algorithms
Techno-functional and nutritional evaluation of Solein single-cell protein and its application in non-dairy yoghurt alternatives
No full textThe growing global population and climate crisis demand expanding non-animal protein options. Single-cell protein biomass, referred to as “Solein”, is produced by the hydrogen-oxidising bacterium Xanthobacter sp. SoF1 and is a promising, sustainable source of protein and dietary fibre, especially when created using renewable energy. This study investigates Solein protein powder (SPP) for its composition and techno-functional properties, comparing it to pea protein isolate (PPI). SPP had a lower fat content and higher dietary fibre, while matching the protein content of PPI. SPP met all indispensable amino acid requirements for adults over the age of three, as outlined by the FAO in 2013.
A milk alternative resembling semi-skimmed cow's milk was produced from SPP and PPI. These emulsions were fermented with a commercial starter culture containing Streptococcus thermophilus. The fermentation process was monitored by tracking pH, total titratable acidity, and microbial growth. The resulting yoghurt alternative (YA) underwent textural and rheological analysis. Solein protein powder yoghurt alternative (SPP-YA) exhibited faster acidification, greater microbial growth, improved water retention, and a texture similar to dairy yoghurt.
Static in vitro digestion revealed moderate protein digestibility of the non-fermented SPP emulsion (63.8–67.5%), based on total amino acids, free amino groups, and total nitrogen, with an in vitro Digestible Indispensable Amino Acid Score (DIAAS) of (51.0 ± 6.1%). Fermentation slightly reduced digestibility (57.8–59.6%) and DIAAS (48.3 ± 1.4%), with isoleucine as the limiting amino acid. This work provides the first insight into the structural and nutritional performance of hydrogen-oxidising bacterial protein in non-dairy YA