Lie symmetries in cortical inspired CNNs

Our scope in this thesis is to propose architectures of CNNs in such a way to model the early visual pathway, including the Lateral Geniculate Nucleus and the Horizontal Connectivity of the primary visual cortex. Moreover, we will show how cortically inspired architectures allow to perform contrast perceptual invariance as well as grouping and the emergence of visual percepts. Particularly, the LGN is modeled with a first layer l0 containing a single filter Ψ0 that pre-filters the image I. Since the RPs of the LGN cells can be modeled as a LoG, we expect to obtain a radially symmetric filter with a similar shape; to this end, we prove the rotational invariance of Ψ0 and we study the influence of this filter to the subsequent layer. Indeed, we compare the statistic distribution of the filters in the second layer l1 of our architecture with the statistic distribution of the RPs of V1 cells of a macaque. Then, we model the horizontal connectivity of V1 implementing a transition kernel K1 to the layer l1. In this setting, we study the vector fields and the association fields induced by the connectivity kernel K1. To this end, we first approximate the filters bank in l1 with a Gabor function and use the parameters just found to re-parameterize the kernel. Thanks to this step, the kernel is now re-parameterized into a sub-Riemmanian space R2 × S1. Now we are able to compare the vector and association fields induced by K1 with the models of the horizontal connectivity.

Heat kernel and worldline path integrals for the Proca theory

In this thesis we study the heat kernel, a useful tool to analyze various properties of different quantum field theories. In particular, we focus on the study of the one-loop effective action and the application of worldline path integrals to derive perturbatively the heat kernel coefficients for the Proca theory of massive vector fields. It turns out that the worldline path integral method encounters some difficulties if the differential operator of the heat kernel is of non-minimal kind. More precisely, a direct recasting of the differential operator in terms of worldline path integrals, produces in the classical action a non-perturbative vertex and the path integral cannot be solved. In this work we wish to find ways to circumvent this issue and to give a suggestion to solve similar problems in other contexts.

Lipschitz regularity for weak solutions of parabolic p-Laplacian type equations in certain subriemannian structures

The main aim of the thesis is to prove the local Lipschitz regularity of the weak solutions to a class of parabolic PDEs modeled on the parabolic p-Laplacian. This result is well known in the Euclidean case and recently has been extended in the Heisenberg group, while higher regularity results are not known in subriemannian parabolic setting. In this thesis we will consider vector fields more general than those in the Heisenberg setting, introducing some technical difficulties. To obtain our main result we will use a Moser-like iteration. Due to the non linearity of the equation, we replace the usual parabolic cylinders with new ones, whose dimension also depends on the L^p norm of the solution. In addition, we deeply simplify the iterative procedure, using the standard Sobolev inequality, instead of the parabolic one.

Learning kernels for support vector machines with polynomial powers of sigmoid

In the pattern recognition research field, Support Vector Machines (SVM) have been an effectiveness tool for classification purposes, being successively employed in many applications. The SVM input data is transformed into a high dimensional space using some kernel functions where linear separation is more likely. However, there are some computational drawbacks associated to SVM. One of them is the computational burden required to find out the more adequate parameters for the kernel mapping considering each non-linearly separable input data space, which reflects the performance of SVM. This paper introduces the Polynomial Powers of Sigmoid for SVM kernel mapping, and it shows their advantages over well-known kernel functions using real and synthetic datasets.

Polynomial and Poisson dependence in free Poisson algebras and free Poisson fields

We prove that any two Poisson dependent elements in a free Poisson algebra and a free Poisson field of characteristic zero are algebraically dependent, thus answering positively a question from Makar-Limanov and Umirbaev (2007) [8]. We apply this result to give a new proof of the tameness of automorphisms for free Poisson algebras of rank two (see Makar-Limanov and Umirbaev (2011) [9], Makar-Limanov et al. (2009) [10]). (C) 2011 Elsevier Inc. All rights reserved.

Polynomial Automorphisms Over Finite Fields

It is shown that the invertible polynomial maps over a finite field Fq , if looked at as bijections Fn,q −→ Fn,q , give all possible bijections in the case q = 2, or q = p^r where p > 2. In the case q = 2^r where r > 1 it is shown that the tame subgroup of the invertible polynomial maps gives only the even bijections, i.e. only half the bijections. As a consequence it is shown that a set S ⊂ Fn,q can be a zero set of a coordinate if and only if #S = q^(n−1).

Polynomial automorphisms over finite fields: Mimicking tame maps by the Derksen group

2010 Mathematics Subject Classification: 14L99, 14R10, 20B27.

Dispersive analysis of KL mu 3 and KLe3 scalar and vector form factors using KTeV data

Using the published KTeV samples of K(L) -> pi(+/-)e(-/+)nu and K(L) -> pi(+/-)mu(-/+)nu decays, we perform a reanalysis of the scalar and vector form factors based on the dispersive parametrization. We obtain phase-space integrals I(K)(e) = 0.15446 +/- 0.00025 and I(K)(mu) = 0.10219 +/- 0.00025. For the scalar form factor parametrization, the only free parameter is the normalized form factor value at the Callan-Treiman point (C); our best-fit results in InC = 0.1915 +/- 0.0122. We also study the sensitivity of C to different parametrizations of the vector form factor. The results for the phase-space integrals and C are then used to make tests of the standard model. Finally, we compare our results with lattice QCD calculations of F(K)/F(pi) and f(+)(0).

Heat flow for Yang-Mills-Higgs fields, part I

The Yang-Mills-Higgs field generalizes the Yang-Mills field. The authors establish the local existence and uniqueness of the weak solution to the heat flow for the Yang-Mills-Higgs field in a vector bundle over a compact Riemannian 4-manifold, and show that the weak solution is gauge-equivalent to a smooth solution and there are at most finite singularities at the maximum existing time.

Enhancement or modulation of the vector competence of ochlerotatus vigilax (Diptera : Culicidae) for Ross River virus by temperature

Two different doses of Ross River virus (1111) were fed to Ochlerotatus vigilax (Skuse), the primary coastal vector in Australia; and blood engorged females were held at different temperatures up to 35 d. After ingesting 10(4.3) CCID50/Mosquito, mosquitoes reared at 18 and 25degreesC (and held at the same temperature) had higher body remnant and head and salivary gland titers than those held at 32degreesC, although infection rates were comparable. At 18, 25, and 32degreesC, respectively, virus was first detected in the salivary glands on days 3, 2, and 3. Based on a previously demonstrated 98.7% concordance between salivary gland infection and transmission, the extrinsic incubation periods were estimated as 5, 4, and 3 d, respectively, for these three temperatures. When Oc. vigilax reared at 18, 25, or 32degreesC were fed a lower dosage of 10(3.3) CCID50 RR/mosquito, and assayed after 7 d extrinsic incubation at these (or combinations of these) temperatures, infection rates and titers were similar. However, by 14 d, infection rates and titers of those reared and held at 18 and 32degreesC were significantly higher and lower, respectively. However, this process was reversible when the moderate 25degreesC was involved, and intermediate infection rates and titers resulted. These data indicate that for the strains of RR and Oc. vigilax used, rearing temperature is unimportant to vector competence in the field, and that ambient temperature variations will modulate or enhance detectable infection rates only after 7 d: extrinsic incubation. Because of the short duration of extrinsic incubation, however, this will do little to influence RR epidemiology, because by this time some Oc. vigilax could be seeking their third blood meal, the latter two being infectious.

Mean anisotropy of homogeneous Gaussian random fields and anisotropic norms of linear translation-invariant operators on multidimensional integer lattices

Sensitivity of output of a linear operator to its input can be quantified in various ways. In Control Theory, the input is usually interpreted as disturbance and the output is to be minimized in some sense. In stochastic worst-case design settings, the disturbance is considered random with imprecisely known probability distribution. The prior set of probability measures can be chosen so as to quantify how far the disturbance deviates from the white-noise hypothesis of Linear Quadratic Gaussian control. Such deviation can be measured by the minimal Kullback-Leibler informational divergence from the Gaussian distributions with zero mean and scalar covariance matrices. The resulting anisotropy functional is defined for finite power random vectors. Originally, anisotropy was introduced for directionally generic random vectors as the relative entropy of the normalized vector with respect to the uniform distribution on the unit sphere. The associated a-anisotropic norm of a matrix is then its maximum root mean square or average energy gain with respect to finite power or directionally generic inputs whose anisotropy is bounded above by a≥0. We give a systematic comparison of the anisotropy functionals and the associated norms. These are considered for unboundedly growing fragments of homogeneous Gaussian random fields on multidimensional integer lattice to yield mean anisotropy. Correspondingly, the anisotropic norms of finite matrices are extended to bounded linear translation invariant operators over such fields.

Reversible current power supply for fast-field cycling nuclear magnetic resonance equipment

The Fast Field-Cycling Nuclear Magnetic Resonance (FFC-NMR) is a technique used to study the molecular dynamics of different types of materials. The main elements of this equipment are a magnet and its power supply. The magnet used as reference in this work is basically a ferromagnetic core with two sets of coils and an air-gap where the materials' sample is placed. The power supply should supply the magnet being the magnet current controlled in order to perform cycles. One of the technical issues of this type of solution is the compensation of the non-linearities associated to the magnetic characteristic of the magnet and to parasitic magnetic fields. To overcome this problem, this paper describes and discusses a solution for the FFC-NMR power supply based on a four quadrant DC/DC converter.

Solar Intensity Forecasting using Artificial Neural Networks and Support Vector Machines

This paper presents several forecasting methodologies based on the application of Artificial Neural Networks (ANN) and Support Vector Machines (SVM), directed to the prediction of the solar radiance intensity. The methodologies differ from each other by using different information in the training of the methods, i.e, different environmental complementary fields such as the wind speed, temperature, and humidity. Additionally, different ways of considering the data series information have been considered. Sensitivity testing has been performed on all methodologies in order to achieve the best parameterizations for the proposed approaches. Results show that the SVM approach using the exponential Radial Basis Function (eRBF) is capable of achieving the best forecasting results, and in half execution time of the ANN based approaches.

Polysaccharide-based freestanding multilayered membranes exhibiting reversible switchable properties

The design of self-standing multilayered structures based on biopolymers has been attracting increasing interest due to their potential in the biomedical field. However, their use has been limited due to their gel-like properties. Herein, we report the combination of covalent and ionic cross-linking, using natural and non-cytotoxic cross-linkers, such as genipin and calcium chloride (CaCl2). Combining both cross-linking types the mechanical properties of the multilayers increased and the water uptake ability decreased. The ionic cross-linking of multilayered chitosan (CHI)â alginate (ALG) films led to freestanding membranes with multiple interesting properties, such as: improved mechanical strength, calcium-induced adhesion and shape memory ability. The use of CaCl2 also offered the possibility of reversibly switching all of these properties by simple immersion in a chelate solution. We attribute the switch-ability of the mechanical properties, shape memory ability and the propensity for induced-adhesion to the ionic cross-linking of the multilayers. These findings suggested the potential of the developed polysaccharide freestanding membranes in a plethora of research fields, including in biomedical and biotechnological fields.

Chebyshev polynomial approximation to approximate partial differential equations

This paper suggests a simple method based on Chebyshev approximation at Chebyshev nodes to approximate partial differential equations. The methodology simply consists in determining the value function by using a set of nodes and basis functions. We provide two examples. Pricing an European option and determining the best policy for chatting down a machinery. The suggested method is flexible, easy to program and efficient. It is also applicable in other fields, providing efficient solutions to complex systems of partial differential equations.