School of Mathematics
http://hdl.handle.net/2262/1
School of Mathematics2017-08-19T18:50:57ZCₒ(X)-structure in C*-algebras, multiplier algebras and tensor products
http://hdl.handle.net/2262/80336
Cₒ(X)-structure in C*-algebras, multiplier algebras and tensor products
McConnell, David
We begin in Chapter 2 with an introduction to the various notions of a bundle of C*-algebras that have appeared throughout the literature, and clarify the definitions of upper- and lower-semicontinuous C*-bundles not explicitly defined in a formal way elsewhere. The definition of C0(X)-algebra, introduced by Kasparov [38], and its relation to C*-bundles is discussed in this chapter also. The purpose of this chapter is to bring together concepts that we will refer to in subsequent sections and which are described using various notations by different authors. Most of this is implicitly understood elsewhere, though Theorem 2.3.12, relating sub-modules of C0(X)-modules and subbundles of C*-bundles, is a new result.
2015-01-01T00:00:00ZPeturbative study of the Chirally Rotated Schrödinger Functionality in Lattice QCD
http://hdl.handle.net/2262/80311
Peturbative study of the Chirally Rotated Schrödinger Functionality in Lattice QCD
Mainar, Pol Vilascea
In this thesis we study the renormalisation and O(a) improvement of the Chirally Rotated Schrödinger Functional (xSF) in perturbation theory. The xSF was originally proposed in [1] as a way of rehabiliting the mechanism of automatic O(a) improvement in the Schrödinger Functional formulation. In order to achieve this in the interacting theory, the finite coefficient of a dimension 3 boundary counterterm has to be tuned. After this, O(a) effects originating from the bulk action or from insertions of composite operators in the bulk will be absent in physical quantities. As in any lattice regularization with SF boundary conditions, extra O(a) effects arise from the boundaries and are cancelled by tuning two dimension 4 boundary coefficients.
2014-01-01T00:00:00ZParameter estimation on neuron models
http://hdl.handle.net/2262/80301
Parameter estimation on neuron models
Lynch, Eoin
Spiking neuron models can accurately predict the response of neurons to somatically injected currents. However, finding parameters for neuronal models that accurately replicate the responses seen in in-vivo experiments is a challenging problem in computational neuroscience. In this thesis, an algorithm is developed for parameter estimation of spiking neuron models from either in-vitro or in-vivo electrophysiological data. The algorithm is a hybrid genetic algorithm which uses a filter based spike train metric, the van Rossum distance, as a fitness function.
2015-01-01T00:00:00ZSpectral functions from lattice QCD at finite temperature
http://hdl.handle.net/2262/80061
Spectral functions from lattice QCD at finite temperature
Harris, Tim
An investigation of the bottonionium spectrum above and below the QCD deconfinement crossover temperature, Tc, was performed using a non-relativistic treatment of the heavy quark on anisotropic lattices with Nf = 2 + 1 flavours of Wilson-clover fermion and a Symanzik-improved gauge action. The spectral functions were reconstructed from the Euclidean correlators using two Bayesian methods to tackle the ill-posed inverse problem, known as the Maximum Entropy Method (MEM) and the Bayesian Reconstruction (BR) method.
2015-01-01T00:00:00Z