41199 research outputs found
Sort by
Utilizzo di autoencoders variazionali per l'identificazione di parametri d'ordine in meccanica statistica
Get PDFIn fisica, molto spesso, si ha a che fare con dati ad alta dimensionalità, per esempio relativi alle coordinate di un sistema atomico o allo spazio delle configurazioni di un sistema di spin. In questi casi, tecniche recentemente introdotte di unsupervised machine learning costituiscono un prezioso aiuto per identificare le cosiddette variabili latenti di un sistema, ossia un ridotto numero di quantità che racchiudono le caratteristiche fondamentali del sistema stesso. Questa riduzione dimensionale si può ottenere con i variational autoencoders (VAE) una particolare forma di rete neurale che cerca di apprendere la distribuzione di probabilità dei dati disponibili. Il lavoro di questa tesi consiste nella costruzione di un VAE per identificare in un modello meccanico statistico una o più variabili latenti che si comportino come parametro d'ordine del sistema. Si cercherà, quindi, di caratterizzare la transizione di fase di questo sistema, allenando un VAE su un dataset opportunamente simulato tramite un codice Monte Carl
Analysis of atmospheric pressure plasma for biological applications
Get PDFIn the last 25 years the field of plasma medicine encountered a large growth. Different sources have been developed and studied, for different purposes: among them, in the Padova laboratories the Plasma Coagulation Controller has been ideated, designed and built. The biological effects of this helium-driven source have been observed both in antibacterial applications and in blood coagulation. As for other sources, the interaction between plasmas and biological substrates which leads to observed effects is a complex combination of different phenomena and actors; the totality of the processes is still far to be understood.
Reactive oxygen and nitrogen species, generated in the interaction of plasma with atmosphere, are considered to have the main role in medical effects of plasmas: the aim of this thesis is to characterize the production of these compounds in the source, and its variation as a function of the power and environmental conditions.
The species are measured using Fourier-transform infrared absorption spectroscopy, a technique which allows a precise and fast measurement of gases concentrations. The source is therefore adapted to the optical system, providing it with all the instruments necessary to control the atmosphere in which the plasma is generated; then, all the setup is optimized, minimizing the errors and the noise sources.
As final results, the presence of ozone, nitrous oxide and nitric acid has been detected. Moreover, the variation in production of these species as function of the atmosphere humidity has been measured, proving that as the humidity rise, the production of ozone fall. Finally, the measured data are compared to theoretical models, finding compatibility
Opacity effects on the evolution of massive stars
No full textAn in-depth analysis of the stellar matter resistance to energy transport, the opacity, in the
context of massive stars with high Zero Age Main Sequence (ZAMS) mass is performed. The opacity of
astrophysical plasmas stands up as key in Stellar Evolution Codes (SECs) input physics, as it is know to
dramatically affect model predictions of structural as well as evolutionary properties of stellar objects; by
virtue of this paramount role, stellar plasma’s opaqueness is examined both as a stand-alone micro-
physical process and as the possible origin of consistent evolutionary effects.
The implementation, in the PAdova TRieste Evolutionary Code (PARSEC), of a refined sub-routine for
relativistic electron scattering, with arbitrary degree of degeneracy, is followed by a thorough discussion
of its consequent evolutionary impacts, within a varied stellar models grid. Furthermore, the author argues
about the importance of thermal conductivity in the most advanced evolutionary stages, thus proceeding
with its inclusion in the code at the highest temperature range. Lastly, the work tests a new prescription,
for atomic and molecular transitions, still in preparation at the Padova group; model tracks of selected
ZAMS from the grid are comprehensively described, offering great insights on the cooler temperatures’
opacity effects on massive stars evolution
Study of performances for Restricted Boltzmann Machines
Get PDFRestricted Boltzmann Machines (RBMs) are one of the most relevant unsupervised learning methods. The aim of this thesis is to study their performances as a function of their parameters. First, we consider binary-valued RBMs and then we introduce the so-called centering trick, which is known to solve the absence of invariance to flip transformations. Moreover, centering also leads to more accurate models. Then, we discuss RBMs with real-valued units. In particular, we focus on rectified linear units, which are able to achieve better generative performances than binary units
Resource curse and corruption: how natural resources may increase corruption in developing countries
Get PDFDuality and the Weak Gravity Conjecture
Get PDFIn the Einstein-Maxwell theory the Weak Gravity Conjecture follows exactly from the positivity bounds on the scattering amplitudes imposed by locality, Lorentz invariance and unitarity of the S-matrix. However, this is in general not true anymore for theories beyond the pure Einstein-Maxwell. An interesting possibility to restore this equivalence is to require the theory to preserve electromagnetic duality. In this work we explore this idea in the context of the axion-dilaton-Maxwell-Einstein theory. First, we study the duality group of this theory and we analytically derive its duality-preserving, 4-derivatives extension. Then, we determine the subset of values of the higher-order coefficients that realizes the Weak Gravity Conjecture by studying the corrections to the charge-to-mass ratio of a black hole solution of such extended theory. Finally, we compare such constraints with the positivity bounds on the scattering amplitudes in order to check the claimed equivalence between the two requirements
Electron mass effects in the prediction of the muon-electron scattering cross sections
Get PDFAbstract (italiano):
Il momento magnetico anomalo del muone, o “g-2” del muone, è una delle grandezze misurate più precisamente in fisica delle particelle e permette di studiare con grande accuratezza la cosiddetta Quantum Field theory (QFT). La discrepanza di lunga data tra il valore sperimentale e quello teorico del momento magnetico anomalo del muone, aμ, ha portato allo studio attento delle correzioni adroniche poiché sono queste quelle che dominano nel risultato dello SM. Qualche anno fa è stato proposto un nuovo approccio per misurare il contributo adronico al “g-2” del muone, aμHLO, misurando il coupling elettromagnetico Δah(q2), per un quadrimomento trasferito di tipo space-like q2=t<0, attraverso dati di scattering. Lo scattering elastico tra muoni ad alta energia e elettroni atomici è stato identificato come il processo ideale per questa misurazione, portando così alla proposta dell’esperimento MUonE al CERN per determinare Δah(q2) dalla sezione d’urto differenziale dello scattering muone-elettrone. Dal punto di vista teorico, per ottenere un risultato compatibile con quello di MUonE, le correzioni alla sezione d’urto differenziale a leading order (LO), next-to-leading order (NLO) and next-next-to-leading order (NNLO) in QED devono essere considerate, insieme a quelle adroniche a NNLO (l’esperimento MUonE si propone di determinare il contributo adronico a NLO). Finora solo le correzioni a NLO in QED sono state calcolare mentre quelle a NNLO in QED, necessarie per raggiungere l’alta precisione richiesta ma MUonE, non si conoscono ancora. Quindi un nuovo metodo per determinare le correzioni a NNLO in QED, basato sull’espansione per regioni, sarebbe auspicabile. In questa tesi vengono studiati i contributi elettromagnetici, elettrodeboli e adronici al momento magnetico del muone. Si passa poi allo studio della sezione d’urto differenziale per lo scattering muone-elettrone a LO e NLO. Qui le divergenze ultraviolette sono regolarizzate in dimensional regularization e i risultati UV-finiti sono ottenuti nel cosidetto on-shell renormalization scheme. Inoltre il contributo di soft-Bremsstrahlung è introdotto per risolvere le divergenze IR. Infine viene applicato il metodo di espansione per regioni all’ampiezza per lo scattering muone-elettrone in QED a NLO confrontando il risultato con l’espansione di Taylor della sezione d’urto differenziale per lo scattering muone-elettrone.
Abstract (inglese):
The muon anomalous magnetic moment, or muon “g-2”, is one of the most precisely measured quantities in particle physics and allows to test Quantum Field Theory (QFT) in its depth, with unprecedented accuracy. The long-standing discrepancy between the experimental and the SM prediction of the muon anomalous magnetic moment, aμ, has kept the hadronic corrections under close scrutiny for several years. In fact, the hadronic uncertainty dominates that of the SM value and is comparable with the experimental one. A few years ago a new approach has been proposed to determine the leading hadronic contribution to the muon “g-2”, aμHLO, measuring the effective electromagnetic coupling Δah(q2), for space-like squared four-momentum transfers q2=t<0, via scattering data. The elastic scattering of high-energy muons on atomic electrons has been identified as an ideal process for this measurement, leading to the proposal of the MUonE experiment at CERN to extract Δah(q2) from the muon-electron scattering differential cross section. On the theory side, in order to obtain a theoretical result which can be compared with the one measured at MUonE, leading order (LO), next-to-leading order (NLO) and next-next-to-leading order (NNLO) QED corrections to the differential cross section have to be considered, together with the hadronic ones a NNLO (the determination of the hadronic NLO contribution is the goal of the MUonE experiment).
Up to now, only the NLO QED corrections to the differential cross section were computed. The QED corrections at NNLO, crucial to interpret the high precision data of future experiment like MuonE, are not yet known. A new method to compute the QED NNLO correction, based on the expansion by regions approach, would be desirable. In this thesis project it is studied the QED, electroweak and hadronic contributions to the anomalous magnetic moment of the muon. Then it is analyzed the muon-electron scattering differential cross section at LO and NLO. Ultraviolet singularities will be regularized via conventional dimensional regularization and UV-finite results are obtained in the on-shell renormalization scheme. Moreover the soft-Bremsstrahlung contribution is introduced to take care of the infrared divergences. Finally the expansion by regions method is applied to the amplitude for the muon-electron scattering at NLO QED and the result is compared with the Taylor expansion of the exact result for the muon-electron scattering differential cross section
Comparison of experimental and simulated beamlet deflection in a MITICA-like extraction system at BATMAN Upgrade
Get PDFAbstract: In collaboration with ITER, a MITICA-like extraction was installed at the BATMAN Upgrade test
facility in 2021 (BUG-MLE). One main difference to the previous extraction system is the presence of
additional deflection compensation magnets which are designed to suppress the row-wise zig-zag
deflection of the accelerated negative ions caused by the co-extracted electron suppression magnets,
mounted in the second grid of the extraction and acceleration system.
The effect of the deflection correction magnets was characterized in dedicated scans with the available
beam diagnostic tools (Carbon Fiber Composite tile calorimetry and Beam Emission Spectroscopy). Since
the correction magnets are installed only in the upper grid half, the corrected beam can be directly
compared to an uncompensated beam.
The experimental investigations are accompanied by simulations. IBSimu is used to model the beamlet
formation and ion-optics inside the extraction system. The in-house developed BBCNI code is exploited to
track the particles towards the CFC and generate synthetic beamlet emission spectra.
The work focuses on physics interpretation by directly comparing measured and synthetic data. In
particular, after outlining the main theoretical aspects behind the BUG-MLE experiment, a series of
simulations is performed with both of the aforementioned codes, allowing a detailed understanding of the
expected results and providing insight on both the robustness of the horizontal deflection correction and its
impact on the estimated beamlet divergence.
Finally, the experimental outcomes are presented and analysed, from both of the diagnostic tools, and
compared with the simulations in order to evaluate the action of the additional magnets in different
conditions