179,312 research outputs found
A mixed-mode cohesive model accounting for small to large openings transition
No full textThis work addresses the formulation of a new mixed-mode cohesive model, able to handle the transition from small to large openings: the proposed model is an extension of the isotropic damage model formulated in [Confalonieri and Perego, JSSCM, 11-2, 2017] for the simulation of mixed-mode delamination with variable mode-ratio, under the assumption of small relative displacements
Multiscale Molecular Modelling of ATP‐Fueled Supramolecular Polymerisation and Depolymerisation**
No full textRaw research data supporting the publication Perego C. et al., ChemSystemsChem 2021, DOI: https://doi.org/10.1002/syst.20200003
Successful treatment of 15 cases of canine traumatic aural hematoma using autologous platelet rich plasma (PRP)
No full textPlatelet-rich plasma (PRP) derived from whole blood, is characterized by platelet (PLT) concentrations above baseline in a small volume of plasma that can accelerate the healing process [1] by providing elevate concentration of platelet-derived growth factors [2] which can stimulate cell proliferation and decrease of inflammatory reaction. Following the excellent results obtained on a small number of subjects previously reported in a preliminary study [3], in this prospective in vivo study we aimed to describe the clinical efficacy of autologous PRP in the treatment of canine traumatic aural hematoma secondary to otitis externa. After approval by the Ethics Committee of the University of Milan and with the owner informed consent, 20 ml of citrate whole blood were obtained from the cephalic vein of 15 dogs with traumatic aural hematoma of different breeds, 6 males and 9 females, with a age range of 1 -15 years (mean ± DS: 7,4 ± 3,7 years). All subjects had a history of multiple centesis, sometimes associated with injections of cortisone in situ, with subsequent recurrences of hematoma. PRP was produced using a semi-automatic closed system (CPUNT 20, Eltek group, Casale Monferrato, Alessandria, Italy) for veterinary use [4]. The serum-hematic content of the auricular pinna was completely drained using one or more 20G needles (depending of hematoma organization) and, using the same hole of the drainage inlet, the PRP was then injected. The dogs were subjected to weekly follow up for a minimum of 45 days from the first treatment. In case of partial or total recurrence of the aural hematoma of the first follow-up, the treatment was repeated with the same procedure. No dog has been subjected to anaesthesia during the procedures. At D0 the aural hematoma was present for 17 ± 13 days and the mean of the drained serum-hematic content was 23 ± 30 ml. Four dogs had a partially organized aural hematoma. 1.3 ± 0.6 ml of PRP were injected, with a mean concentration of 1185 ± 908 x 103/μl PLT (minimum value: 308 x 103/μl PLT maximum value: 4141 x 103/μl PLT, 500% mean increase compared to whole blood). 12/15 subjects were treated with a single application (Group A1), 3/15 with two applications (Group A2). 2/15 subjects were lost after the first follow up. For the remaining 13/15 the mean healing time was 15.8 ± 8.1 days (A1) and 24 ± 5.2 days (A2). No subjects showed recurrences at 45 days follow up. No side effects have been registered. The in situ administration of PRP was effective in the treatment of traumatic aural hematoma secondaty to otitis externa in dogs, leading to complete resolution of the disease in all treated subjects.
[1] Marx R.E. Platelet-Rich Plasma (PRP), what is PRP and what is not PRP? Implant Dentistry 2001, 10, 225-228. [2] Souza T.F. et al. Healing and expression of growth factors (TGF-β and PDGF) in canine radial ostectomy gap containing platelet-rich plasma. Vet Comp Orthop Trauma, 2012, 25, 445-452. [3] Perego R. et al. Efficacia clinica del plasma ricco di piastrine (prp) autologo ottenuto con metodo chiuso semi-automatico nel trattamento dell’otoematoma nel cane. 2017. 56° Congresso AIVPA, Piacenza. [4] Perego R. et al. Evaluation of a commercial closed system for autologous platelet-rich plasma production in dog. 2016. ESVCP-ESVONC Congress, Nantes
Vocatives in subitles: A survey across genre
No full textThe purpose of this contribution is to investigate the function of vocatives and their translation in interlinguistic subtitles over different film genres. Our previous investigation (Bruti, Perego 2005) was based on a small corpus that included 2 British and 2 American films, belonging approximately to the same genre, i.e. COMEDY (with the exception of "The Talented Mr Ripley", which begins as a COMEDY but turns into a DRAMATIC MYSTERY STORY). This project aims to investigate the various roles vocatives play in the construction of the narrative according to the different needs that different film genres aim to fulfil. The corpus has therefore been extended to include: a full-length animated feature from Walt Disney Pictures ("Bambi", D. Hand, 1942), an action film ("Lethal Weapon 4", R. Donner, 1998), an adaptation from a literary masterpiece ("Sense and Sensibility", A. Lee, 1996), a popular comedy series (two episodes of "Sex and the City", Season 4, “The Agony and the Ex-tasy”, M.P. King, 2001 and “I heart NY”, M.P. King 2002) and an animated series (two episodes of "The Simpsons", “Homer in the night”, R. Moore, 1989-90; “Homer the Moe”, J. Kamerman, 2001-02)
Entity linking on philosophical documents
No full textEntity Linking consists in automatically enriching a document by detecting the text fragments mentioning a given entity in an external knowledge base, e.g., Wikipedia. This problem is a hot research topic due to its impact in several text-understanding related tasks. However, its application to some specfiic, restricted topic domains has not received much attention. In this work we study how we can improve entity linking performance by exploiting a domain-oriented knowledge base, obtained by filtering out from Wikipedia the entities that are not relevant for the target domain. We focus on the philosophical domain, and we experiment a combination of three different entity filtering approaches: one based on the Philosophy" category of Wikipedia, and two based on similarity metrics between philosophical documents and the textual description of the entities in the knowledge base, namely cosine similarity and Kullback-Leibler divergence. We apply traditional entity linking strategies to the domainoriented knowledge base obtained with these filtering techniques. Finally, we use the resulting enriched documents to conduct a preliminary user study with an expert in the area
Audiovisual genre and the translation of vocatives in interlingual subtitles
No full textIn this work we set out to investigate the role of audiovisual genre in the translation of vocatives in interlingual subtitles. By merging the distinct research areas of film genre and audiovisual translation, we mean to shed light on both and to evidence their mutual influences. Our investigation is based on a varied sample of English audiovisual products subtitled in Italian. The corpus consists of an old sample of four British and American films - mainly comedies and dramas (cf. Bruti, Perego 2005) - , i.e., Sliding Doors (P. Howitt, 1998, UK), The Talented Mr Ripley (A. Minghella, 1999, USA-Italy), Shallow Hal (Farrelly Brothers, 2002, USA) and East is East (D. O’Donnel, 1999, UK), and a new one, which includes further material, i.e., a full-length animated feature from Walt Disney Pictures (Bambi, D. Hand, 1942), an action film (Lethal Weapon 4, R. Donner, 1998), an adaptation from a literary masterpiece (Sense and Sensibility, A. Lee, 1996), a popular comedy series (two episodes of Sex and the City, Season 4, “The Agony and the Ex-tasy”, M.P. King, 2001 and “I heart NY”, M.P. King 2002) and an animated series (two episodes of The Simpsons, “Homer in the night”, R. Moore, 1989-90; “Homer the Moe”, J. Kamerman, 2001-02). Although comedy still predominates, the corpus is certainly comparatively more assorted and representative
Beyond accuracy: enhancing multiple perspectives of recommendation through multi-objective optimization and evaluation
No full textI sistemi di raccomandazione sono diventati strumenti essenziali per alleviare il problema dell’information overload, fornendo suggerimenti personalizzati in vari settori, tra cui l'e-commerce, le piattaforme di streaming e i social network. Tradizionalmente, la valutazione e l'ottimizzazione dei sistemi di raccomandazione si sono concentrate sull'accuratezza come parametro principale di successo. Sebbene l'accuratezza sia fondamentale per predire le preferenze degli utenti, essa non affronta dimensioni più ampie, cruciali per migliorare la soddisfazione degli utenti, garantire l'equità degli stakeholder e affrontare gli impatti sociali. Inoltre, quando si considerano più obiettivi, spesso sorgono conflitti, cioè il miglioramento di un obiettivo può influire negativamente sugli altri, portando a uno spettro di possibili soluzioni ottimali. Queste sfide danno origine a diverse domande critiche: Come possono i sistemi di raccomandazione evolversi per bilanciare l'accuratezza con altri obiettivi, come la diversità, la novità e l'equità, soddisfacendo al contempo le esigenze di più parti interessate, tra cui utenti, fornitori di contenuti e piattaforme? Come possiamo valutare simultaneamente l'efficacia dei sistemi di raccomandazione attraverso diversi criteri? Come si può selezionare un'unica soluzione ottimale da un insieme di compromessi? Infine, possiamo progettare un framework generico per l'ottimizzazione dei sistemi di raccomandazione che tenga conto di obiettivi multipli, spesso in conflitto tra loro? Queste domande evidenziano le principali sfide aperte nel campo della ricerca sui sistemi di raccomandazione.
Questa tesi affronta queste lacune concentrandosi su due aree principali: le metodologie per la valutazione multi-obiettivo deli sistemi di raccomandazione e le sfide associate alla progettazione di sistemi di raccomandazione multi-obiettivo. Dopo un'esplorazione approfondita del background dei sistemi di raccomandazione e dell'ottimizzazione multi-obiettivo, la tesi fornisce contributi significativi nelle seguenti aree:
(i) l'applicazione delle frontiere di Pareto per condurre una valutazione multi-obiettivo di sistemi di raccomandazione basati su grafi, concentrandosi sugli aspetti di equità;
(ii) l'introduzione di indicatori di qualità per le frontiere di Pareto per scoprire il potenziale dei sistemi di raccomandazione al di là delle tradizionali metriche di accuratezza;
(iii) lo sviluppo di un quadro analitico per valutare la sensibilità dei sistemi di raccomandazione al tuning degli iperparametri in scenari multi-obiettivo;
(iv) uno studio sulla riproducibilità che identifica le principali sfide e ambiguità nella progettazione e nella valutazione dei sistemi di raccomandazione multi-obiettivo;
(v) la proposta di una nuova strategia di selezione di soluzioni Pareto-ottimali post-hoc, adattata esplicitamente ai task di raccomandazione;
(vi) la progettazione di un framework flessibile di sistema di raccomandazione multi-obiettivo che incorpora loss functions indipendenti dagli obiettivi e consapevoli della loro magnitudine per ottenere l'ottimizzazione di diversi obiettivi di raccomandazione.Recommender Systems (RSs) have become essential tools for alleviating information overload by providing personalized suggestions across various domains, including e-commerce, streaming platforms, and social networks. Traditionally, the evaluation and optimization of RSs have centered on accuracy as the primary success metric. While accuracy is critical for predicting user preferences, it fails to address broader dimensions crucial for enhancing user satisfaction, ensuring stakeholder fairness, and addressing societal impacts. Moreover, when multiple objectives are considered, conflicts often arise, i.e., improving one objective can detrimentally affect others, leading to a spectrum of possible optima. These challenges give rise to several critical questions: How can RSs evolve to balance accuracy with other objectives, such as diversity, novelty, and fairness, while meeting the needs of multiple stakeholders, including users, content providers, and platforms? How can we simultaneously evaluate RS effectiveness across diverse criteria? How can a single optimal solution be selected from a set of trade-offs? Finally, can we design a generic framework for optimizing RSs that accounts for multiple, often conflicting objectives? These questions highlight key open challenges in the field of RS research.
This dissertation addresses these gaps by focusing on two main areas: methodologies for multi-objective evaluation of RSs and the challenges associated with designing Multi-Objective Recommender Systems (MORSs). After an in-depth exploration of the background of RSs and multi-objective optimization, the thesis makes significant contributions in the following areas:
(i) the application of Pareto frontiers to conduct a multi-objective evaluation of graph-based RSs, focusing on fairness aspects;
(ii) the introduction of quality indicators for Pareto frontiers to uncover the potential of RSs beyond traditional accuracy metrics;
(iii) the development of an analytical framework to assess the sensitivity of RSs to hyper-parameter tuning in multi-objective scenarios;
(iv) a reproducibility study that identifies key challenges and ambiguities in the design and evaluation of MORSs;
(v) the proposal of a novel, post-hoc Pareto-optimal solution selection strategy tailored explicitly for RS tasks;
(vi) designing a flexible MORS framework incorporating objective-agnostic and scale-aware loss functions to achieve optimization across diverse recommendation objectives
Searching the Optimal Folding Routes of a Complex Lasso Protein
No full textUnderstanding how polypeptides can efficiently and reproducibly attain a self-entangled conformation is a compelling biophysical challenge that might shed new light on our general knowledge of protein folding. Complex lassos, namely self-entangled protein structures characterized by a covalent loop sealed by a cysteine bridge, represent an ideal test system in the framework of entangled folding. Indeed, because cysteine bridges form in oxidizing conditions, they can be used as on/off switches of the structure topology to investigate the role played by the backbone entanglement in the process. In this work, we have used molecular dynamics to simulate the folding of a complex lasso glycoprotein, granulocyte-macrophage colony-stimulating factor, modeling both reducing and oxidizing conditions. Together with a well-established Gō-like description, we have employed the elastic folder model, a coarse-grained, minimalistic representation of the polypeptide chain driven by a structure-based angular potential. The purpose of this study is to assess the kinetically optimal pathways in relation to the formation of the native topology. To this end, we have implemented an evolutionary strategy that tunes the elastic folder model potentials to maximize the folding probability within the early stages of the dynamics. The resulting protein model is capable of folding with high success rate, avoiding the kinetic traps that hamper the efficient folding in the other tested models. Employing specifically designed topological descriptors, we could observe that the selected folding routes avoid the topological bottleneck by locking the cysteine bridge after the topology is formed. These results provide valuable insights on the selection of mechanisms in self-entangled protein folding while, at the same time, the proposed methodology can complement the usage of established minimalistic models and draw useful guidelines for more detailed simulations
Computational methods in the study of self-entangled proteins: a critical appraisal
No full textThe existence of self-entangled proteins, the native structure of which features a complex topology, unveils puzzling, and thus fascinating, aspects of protein biology and evolution. The discovery that a polypeptide chain can encode the capability to self-entangle in an efficient and reproducible way during folding, has raised many questions, regarding the possible function of these knots, their conservation along evolution, and their role in the folding paradigm. Understanding the function and origin of these entanglements would lead to deep implications in protein science, and this has stimulated the scientific community to investigate self-entangled proteins for decades by now. In this endeavour, advanced experimental techniques are more and more supported by computational approaches, that can provide theoretical guidelines for the interpretation of experimental results, and for the effective design of new experiments. In this review we provide an introduction to the computational study of self-entangled proteins, focusing in particular on the methodological developments related to this research field. A comprehensive collection of techniques is gathered, ranging from knot theory algorithms, that allow detection and classification of protein topology, to Monte Carlo or molecular dynamics strategies, that constitute crucial instruments for investigating thermodynamics and kinetics of this class of proteins
- …
