10081 research outputs found
Sort by
Understanding the causes of junctional failure in lumbar spine surgery through retrospective clinical analyses and ex-vivo tests
No full textSpine diseases and failure have a significant effect on the patient’s quality of life. The aim of this PhD project was to improve the understanding the mechanisms leading to the spine instability and junctional pathology in the lumbar spine after posterior fixation. A combination of ex-vivo experiments and retrospective clinical analyses were performed.
In the first part, cadaver spine segments were sequentially tested in the intact condition, after two-level hemilaminectomy, after a full laminectomy and after posterior fixation. Hemilaminectomy did not significantly affect the biomechanics of the lumbar spine. Conversely, after laminectomy, the range of motion significantly increased in flexion. Lateral bending was the most critical configuration for the strain in the intervertebral discs. Posterior fixation significantly reduced the range of motion and increased the strain in the disc adjacent the fixation both in flexion and in lateral bending.
Two retrospective studies focused on patients with posterior fixation on the lumbar spine who required a revision surgery due to a failure in the caudal region. The junctional pathology onset was assessed on the pre-operative and post-operative spinopelvic parameters both in the sagittal and coronal planes, on the correction performed, on the patients’ clinical data, and addressing the different mechanisms of failure. In the sagittal plane (1690 patients included), the likelihood of mechanical failure was higher in patients older than 40 years with a thoraco-lumbar fixation, where some spinopelvic parameters were not properly restored. In the coronal plane (105 scoliotic patients included), the main parameters and their post-operative variations leading to failure were identified.
In conclusion, the findings of these ex-vivo and clinical analyses could support the surgeons in being aware of possible outcomes when a treatment is needed, and when possible, to personalize spine surgeries, aiming to the long term to decrease the mechanical failure and revision surgeries
A downscaling exercise for the coastal ocean with the perfect model approach: a study for the Adriatic Sea
Get PDFThis PhD thesis proposes a downscaling exercise with the perfect model approach. A large domain high resolution experiment covering the entire Adriatic Sea, referred to as the “truth”, is used as a benchmark for assessing downscaling configurations for a small nested domain, the Northern Adriatic Sea. A set of experimental setup with different nesting strategies have been performed based on the NEMO ocean model, where child experiments are finally compared with the truth reference. The downscaling performance of each strategy was estimated by kinetic energy spectral analysis. Overall, the child experiments were able to regenerate small scale features and recover the energy not present at the boundaries. A downscaling ratio of 1:3 has proven to be suitable for allowing regeneration of kinetic energy in the inner domain. At larger ratios (i.e. 1:5), drawbacks associated with the lateral open boundaries are propagated resulting in overshooting of energy at specific frequencies.
The analyses of local physical processes (dense water formation) and extreme events (marine heat waves) served the scope of judging the capability of reproducing the thermohaline variability of the Northern Adriatic Sea by looking at the effects of local drivers at small scales and/or remote drivers at large scales. The perfect model approach underscores that when remote drivers are missed by the parent model, i.e. the salt water inflow through the Otranto Strait during the cyclonic phase of the Northern Ionian Gyre, the child model may fail the reproducibility of local dense water formation.
A final methodology is proposed based on the spectral nudging technique, proven to be a suitable tool for coastal ocean modelling application. The resulting inner domain circulation evolved with better representation of water inflow at the Otranto Strait and thermohaline properties also at remote shallow coastal areas of the Northern Adriatic subdomain
Knowledge-enhanced natural language processing
Get PDFLanguage is a fundamental part of being human. It allows people to communicate thoughts, express ideas, create memories, and build mutual comprehension. Developing machines capable of understanding language automatically has been a long-standing dream in artificial intelligence.
Over the past decade, dramatic and unprecedented breakthroughs in natural language processing have been driven by transformer-based networks scaled on progressively larger parameter counts and training data.
Large language models have recently become prominent, revealing early signs of general intelligence and ushering in a new era of human-machine interaction.
However, although pre-trained language models could effectively store numerous facts encountered at training time, their knowledge awareness is still far from satisfactory.
Current models grapple with lexical superficiality, human reporting bias, hallucination, poor reasoning, black-box behavior, and computational complexity.
These open issues have led researchers to question the adequacy of scaling as the sole condition for scientific progress.
A growing number of influential thinkers recognize that the key direction for substantial advancements lies in the incorporation of semantic-level abstractions and pertinent context derived from external knowledge sources.
In this dissertation, we study knowledge-enhanced natural language processing, presenting multiple techniques for knowledge extraction, representation, and injection.
We show that semantic parsing, graph representation learning, retrieval augmented generation, multi-modal signals, and back-propagation through combinatorial solvers are integral components of the solution.
Overall, this work represents a significant step toward delineating a new generation of language models that prioritize meaning, reasoning, and interpretability
A numerical study of subsurface oil spills
Get PDFThis thesis addresses numerical modelling of oil entering the ocean from subsurface sources like drilling well blowouts or pipeline failures. We introduce a single-phase oil model, excluding a gaseous component. We've developed a novel Python-based nearfield plume model where buoyant oil forms a coherent plume with a self-similar structure along vertical uplifting. A key contribution is defining essential plume variables and equations for the coupled water-oil system. The nearfield component is validated using lab-scale and real-scale experiments in the North Sea, including a sensitivity analysis on entrainment coefficients, fitting well with in-situ data.
As the plume reaches a terminal level due to loss of momentum and buoyancy from ocean stratification, a farfield stage begins. Here, individual oil droplets disperse via ocean currents and eddy-turbulence. The far-field condition originates from the near-field's final state, where the plume element dissolves into oil parcels. Nearfield oil-water mixture at the terminal level seamlessly connects to horizontal and vertical spreading. The farfield component employs a Lagrangian Particle Tracking model (OceanParcels), incorporating a vertical component into the advection-diffusion problem. Size-dependent buoyancy forms distinct clusters with varying resurfacing times.
The newly developed plume model and the farfield model, both Python-based, are unified under UWORM (UnderWater Oil Release Model). They utilize 3D ocean state data from the Copernicus Marine Service (currents velocity, temperature, and salinity fields). Uncertainties in both models are discussed for future work. In the nearfield phase, uncertainties focus on entrainment parameter choices and input ocean data, while ocean currents and stratification affect the terminal level. In the farfield, uncertainties concern the resolution of input ocean currents and eddy diffusivity parameterization. Additionally, droplet size distribution significantly impacts the total volume of oil resurfaced
Through Medusa’s gaze: exploring women’s perception of gender inequalities and its relevance to social change
Get PDFDespite significant progress toward gender equality, women still face disparities across various life domains, including professional and family contexts. Policymakers now distinguish these domains to measure inequalities more accurately and identify areas needing intervention. However, social psychology has not mirrored this trend, despite its potential to elucidate women’s psychological reactions to perceived inequalities. This dissertation fills this gap by exploring perceived gender inequalities and associated psychological reactions. Chapter 1 reviews theoretical foundations: the social identity approach, the rejection-identification model, and the theory of relative deprivation. Chapters 2 to 6 present empirical studies divided into four sections, examining perceived gender inequalities (Section A), social identity processes and collective action (Section B), professional aspirations (Section C), and attitudes toward minority groups (Section D).
Section A developed a multidimensional conceptualization of perceived gender inequalities, including a pilot qualitative study and five quantitative studies in Italy and the UK (N = 1690) to validate an instrument measuring these perceptions. Section B explored the relationship between perceptions of gender inequalities, social identification with women and feminists, and collective action intentions, considering legitimacy perceptions. This included two correlational studies in Italy and Turkey (N = 976) and an experimental study in the UK (N = 293). Section C examined the impact of perceived inequalities on women’s career aspirations, considering contingencies of self-worth, through two correlational studies among university students in Italy and Spain (N = 696). Section D investigated the relationship between perceived inequalities and attitudes toward other minorities, using a correlational study among Italians (N = 493) to assess attitudes towards gays and lesbians, trans women, and migrants.
Chapter 7 concludes by discussing findings, theoretical and practical implications, limitations, and future research directions. This work underscores the need for a multidimensional understanding of gender inequalities to create awareness and reduce these disparities
The digital evolution of parody. The purpose of online parody in copyright law and freedom of expression.
Get PDFThe thesis aims to focus on parody which has always been tied to art, marked by stinging irony. This form of artistic expression is gaining significant relevance on today’s internet, where, thanks to technological development, it actively facilitates peoples’ participation in the information society. The advent of digital technology has renewed the ancient parodic technique, so that the online parody is perfectly embedded in today's society, with a great many sectors using messages that are based on parodic rhetoric. Online parody presents peculiarities of online interactions, such as the speed and nature of online artistic productions, and problems that cannot always be engaged with using traditional legal theories and categories. The background of the legal analysis is divided into private and public interests. The dissertation will present the reconstruction of parody exception within European Directives on Copyright law. This part of the thesis will study how and when the law has predicted some kind of protection for parody and what are the differences for the regulation of this art form between a well-known model in the offline world and a still unsure model in the online world. In this regard, online parody will be described as a symbol for a shared digital economy of the creativity, aligning with the necessary rethinking of copyright law. The thesis will also dedicate on the study of the critical spirit inside the Internet. Freedom of parody will be pointing out the public interest, the relevant fundamental rights and the necessary limitations to which any expression is subject. The outline will include the analysis of the content-sharing providers and their role in regulating user-generated parody; the social media are get used to intervene in online parody through policies aimed both at contrasting hate speech and fake news and at maintaining a free and democratic platform
From dark matter to dark sectors: exploring signatures at multiple scales
Get PDFThe quest to understand the nature of dark matter is one of the most compelling challenges in modern physics. This dissertation investigates various dark matter models, aiming to establish connections between theory and observations. Beginning with indirect detection searches, we focus on dark matter annihilation into monochromatic photon lines. We examine the potential of this signature by constraining a top-philic dark matter model and the inert doublet model, using Fermi-LAT and HESS experimental data. We explore the singlet scalar Higgs portal model on the resonance region, combining data from the galactic center, dwarf galaxies and antiprotons. The model accommodates these measurements along with the relic density, while being unchallenged by current direct detection and collider constraints. We also investigate the phenomenology of superheavy decaying dark matter, constraining its lifetime on the basis of gamma-rays and future neutrino observations. Furthermore, we introduce light dark sectors, focusing on Standard Model extensions featuring a vector portal, involving a new dark gauge symmetry with a kinetically mixed dark photon. In particular, we focus on light dark photons, proposing models containing multiple dark fermions, generalizing the idea of the inelastic dark matter. We probe these rich dark sector models by searching for semi-visible dark photon decays, where electron-positron pairs are produced alongside missing energy. This decay pattern allows recasting and relaxing the main constraints, opening new regions of the parameter space, that can account for the anomalous magnetic moment of the muon, and be tested in the future. Finally, we present a search for electron-positron pairs produced within dark sector models in the MicroBooNE experiment, where heavier dark fermions are produced by beam neutrinos upscattering, further decaying to electron-positron pairs. This search has the potential to probe the parameter space that could provide an explanation for the MiniBooNE low-energy excess
De privati iudicis officio. Contribution to the study of attribution criterion of a judge's liability in cases of litem suam facere
Get PDFLa presente ricerca esamina il criterio di attribuzione della responsabilità del giudice, concentrandosi sulla fattispecie del iudex qui litem suam fecit nell'antica Roma. Le fonti suggeriscono che i giudici potessero essere ritenuti responsabili per violazioni procedurali, anticipando il concetto di litem suam facere, come si evince da XII Tab. 2.2. Fonti del II secolo a.C. identificano il litem suam facere come l'assenza ingiustificata del giudice dall'udienza (Mac. Sat. 16,16,15; Gell. Noct. Att. 10,4-5). Altre fonti confermano la natura oggettiva della responsabilità del giudice: P.Ant. I,22 (recto) e lex Irn. cap. 91 puniscono il giudice si neque diffidit neque sententiam dixit; Gaio (4,52) afferma che il giudice fa propria la lite se viola le istruzioni della formula. Tuttavia, la linea ricostruttiva del litem suam facere è ostacolata dalla presenza del sintagma licet per imprudentia (D. 44,7,5,4; D. 50,13,6; I. 4,5 pr.) e da dolus malus in fraudem legis (D. 5,1,15,1). Queste indicazioni non modificano la natura oggettiva della responsabilità del giudice. Si ammette, all'interno del campo di applicazione del litem suam facere di natura oggettiva, anche casi che sarebbero stati giustificati (imprudentia) o comportamenti che altrimenti non sarebbero stati puniti (dolus in fraudem legis). La concezione colposa del litem suam facere è attribuita a Teofilo, che ha influenzato gli scoliasti e i giuristi moderni. Tuttavia, nel diritto romano classico, la responsabilità non richiedeva un criterio di attribuzione soggettivo. Solo nel VI secolo d.C. è emersa la necessità della colpevolezza, portando a un malinteso sulla natura soggettiva della responsabilità del giudice che fa sua la lite.The research examines the criterion of attribution of judge’s liability, focusing on the concept of iudex qui litem suam fecit in ancient Rome. Sources suggest that judges could be held liable for procedural violations, foreshadowing the concept of litem suam facere, as seen in XII Tab. 2.2. Sources from the 2nd century BC identify litem suam facere as the unjustified absence of the judge from the hearing (Mac. Sat. 16,16,15; Gell. Noct. Att. 10,4-5). Other sources confirm the objective nature of judge’s liability: P.Ant. I,22 (recto) and lex Irn. cap. 91 punish the judge si neque diffidit neque sententiam dixit; Gaius (4,52) affirms that the judge litem suam fecit if he violates the instructions of the formula. However, the reconstructive line of litem suam facere is hindered by licet per imprudentia (D. 44,7,5,4; D. 50,13,6; I. 4,5 pr.) and by dolus malus in fraudem legis (D. 5,1,15,1). These indications, however, do not change the objective nature of judge's liability. It admits within the scope of objective litem suam facere even cases that would have been justified (imprudentia) or would not have been punished (dolus in fraudem legis). The negligent conception of litem suam facere is attributed to Theophilus, who influenced subsequent scholastics and modern jurists. However, in classical Roman law, responsibility did not require a subjective attribution criterion. Only in the 6th century AD did the need for culpability arise, leading to a misunderstanding about the subjective nature of judge’s liability
Characterizing “Exposure-PK/PD-Biomarker” relationships in patients with hematologic malignancies
Get PDFAntimicrobial stewardship programs are gaining more and more relevance in optimizing anti-infective treatment and in preventing the emergence of antimicrobial resistance. Personalization of antimicrobial treatment based on real-time therapeutic drug morning (TDM) and dosing adaptation may represent an important tool in antimicrobial stewardship programs.
In this Ph.D project, we aim to focus on differences in pharmacokinetics (PK) for meropenem and piperacillin/tazobactam and host response biomarkers (e.g., C-reactive protein) in severe Gram‐negative related infections occurring in oncohematologic patients. We are interested in identifying optimized model‐based individualized dosing strategies for these antibiotics focusing on biomarkers-guided prediction of PK and pharmacodynamic (PD) parameters using population PK/PD modelling. We expect to identify optimal model‐based dosing targets for these antibiotics for special populations for implementation in TDM routines, and mathematical models characterizing the relationship between biomarkers and outcomes in these populations
Vehicle routing and location routing problems with minimum latency: algorithms and models
Get PDFLatency can be defined as the sum of the arrival times at the customers.
Minimum latency problems are specially relevant in applications related to humanitarian logistics.
This thesis presents algorithms for solving a family of vehicle routing problems with minimum latency.
First the latency location routing problem (LLRP) is considered. It consists of determining the subset of depots to be opened, and the routes that a set of homogeneous capacitated vehicles must perform in order to visit a set of customers such that the sum of the demands of the customers assigned to each vehicle does not exceed the capacity of the vehicle.
For solving this problem three metaheuristic algorithms combining simulated annealing and variable neighborhood descent, and an iterated local search (ILS) algorithm, are proposed.
Furthermore, the multi-depot cumulative capacitated vehicle routing problem (MDCCVRP) and the multi-depot k-traveling repairman problem (MDk-TRP) are solved with the proposed ILS algorithm. The MDCCVRP is a special case of the LLRP in which all the depots can be opened, and the MDk-TRP is a special case of the MDCCVRP in which the capacity constraints are relaxed.
Finally, a LLRP with stochastic travel times is studied.
A two-stage stochastic programming model and a variable neighborhood search algorithm are proposed for solving the problem.
Furthermore a sampling method is developed for tackling instances with an infinite number of scenarios.
Extensive computational experiments show that the proposed methods are effective for solving the problems under study