This thesis is devoted to creating a systematic way of determining all inequivalent orthogonal coordinate systems which separate the Hamilton-Jacobi equation for a given natural Hamiltonian defined on three-dimensional spaces of constant, non-zero curvature. To achieve this, we represent the problem with Killing tensors and employ the recently developed invariant theory of Killing tensors. Killing tensors on the model spaces of spherical and hyperbolic space enjoy a remarkably simple form; even more striking is the fact that their parameter tensors admit the same symmetries as the Riemann curvature tensor, and thus can be considered algebraic curvature tensors. Using this property to obtain invariants and covariants of Killing tensors, together with the web symmetries of the associated orthogonal coordinate webs, we establish an equivalence criterion for each space. In the case of three-dimensional spherical space, we demonstrate the surprising result that these webs can be distinguished purely by the symmetries of the web. In the case of three-dimensional hyperbolic space, we use a combination of web symmetries, invariants and covariants to achieve an equivalence criterion. To completely solve the equivalence problem in each case, we develop a method for determining the moving frame map for an arbitrary Killing tensor of the space. This is achieved by defining an algebraic Ricci tensor. Solutions to equivalence problems of Killing tensors are particularly useful in the areas of multiseparability and superintegrability. This is evidenced by our analysis of symmetric potentials defined on three-dimensional spherical and hyperbolic space. Using the most general Killing tensor of a symmetry subspace, we derive the most general potential “compatible” with this Killing tensor. As a further example, we introduce the notion of a joint invariant in the vector space of Killing tensors and use them to characterize a well-known superintegrable potential in the plane. xiii

Neste trabalho mostramos que dado um domínio exterior de classe C0 contido em uma superfície umb lica de H3; com curvatura média constante H 2 [0; 1); existe uma família de gracos de Killing com curvatura média constante H: O bordo de cada um destes gracos está contido nesta superfície umbílica e a norma do gradiente da função no bordo pode ser prescrita por um certo valor s 0. / In this paper we show that given an exterior domain of class C0 contained in an umbilical surface of H3; with constant mean curvature H 2 [0; 1); there exists a family of Killing graphs with constant mean curvature H: The boundary of each of these graphs is contained in this umbilical surface and the norm of the gradient of the function in the boundary can be prescribed by a certain value s 0:

Neste trabalho mostramos que dado um domínio exterior de classe C0 contido em uma superfície umb lica de H3; com curvatura média constante H 2 [0; 1); existe uma família de gracos de Killing com curvatura média constante H: O bordo de cada um destes gracos está contido nesta superfície umbílica e a norma do gradiente da função no bordo pode ser prescrita por um certo valor s 0. / In this paper we show that given an exterior domain of class C0 contained in an umbilical surface of H3; with constant mean curvature H 2 [0; 1); there exists a family of Killing graphs with constant mean curvature H: The boundary of each of these graphs is contained in this umbilical surface and the norm of the gradient of the function in the boundary can be prescribed by a certain value s 0:

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / O propÃsito desta tese à obter teoremas de caracterizaÃÃo de hipersuperfÃcies tipo-espaÃo completas isometricamente imersas num ambiente semi-Riemanniano mediante alguma restriÃÃo sobre a aplicaÃÃo de gauss ou sobre as r-curvaturas mÃdias destes objetos. Iniciamos nosso trabalho dando condiÃÃes necessÃrias para garantir a umbilicidade de hipersuperfÃcies imersas no espaÃo hiperbÃlico Hn+1 com aplicaÃÃo de Gauss prescrita. Em seguida, obtemos alguns resultados de unicidade de hipersuperfÃcies completas com curvaturas de ordem superior limitadas num ambiente do tipo et x et Mn supondo uma restriÃÃo apropriada sobre o Ãngulo normal da hipersuperfÃcie em questÃo. Na Ãltima parte deste trabalho, obtemos resultados tipo-Bernstein considerando grÃficos verticais completos com curvatura mÃdia constante imersos num produto warped Riemanniano I xf Mn onde supomos uma conhecida condiÃÃo de convergÃncia sobre a curvatura seccional da fibra Mn. / The purpose of this thesis is to obtain characterization theorems of complete spacelike hypersurfaces isometrically immersed in a semi-Riemannian ambient space under some restrictions on the Gauss mapping or about the r-mean curvatures of these objects. We start our work by providing necessary conditions to ensure the umbilicity of immersed hypersurfaces in the hyperbolic space Hn+1 with prescribed Gauss mapping. Next, we obtain some uniqueness results of complete hypersurfaces with bounded higher order mean curvatures in a space ER x et Mn where we suppose an appropriate condition on the normal angle of the hypersurface. In the last part of this work, we obtain Bernstein-type results concerning to complete vertical graphs with constant mean curvature immersed in a Riemannian warped product I x f Mn, where we suppose a well know convergence condition on the sectional curvature of the fibre Mn.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / Apresentamos um mÃtodo para a obtenÃÃo de limites inferiores para o primeiro autovalor de Dirichlet em termos de campos vetoriais com divergÃncia positiva. Aplicando-o ao gradiente de uma funÃÃo distante, obtemos estimativas de de autovalor em bolas geodÃsicas em cut locus e dos domÃnios de subvariedades com curvatura mÃdia localmente limitada.Para subvariedades das variedade de Hadamard com limites mÃdios de curvaturas, estes limites inferiores dependem da dimensÃo das subvariedades e limite sobre sua curvatura mÃdia. / We present a method to obtain lower bounds for first Dirichlet eigenvalue in terms of vector fields with positive divergence. Applying this to the gradient of a distance function we obtain estimates of eigenvalue of geodesic balls inside the cut locus and of domains in submanifolds with locally bounded mean curvature. For submanifolds of Hadamard manifolds with bounded mean curvature these lower bounds depend only on the dimension of the submanifold and the bound on its mean curvature.

[pt] Um resultado clássico no âmbito de equações diferenciais parciais e de geometria diferencial é o seguinte: Dada uma constante a existe uma condição da fronteira do domínio (Omega) de maneira que o problema de Dirichlet para a equação da curvatura média a no espaço Euclidiano é sempre solúvel. Este é um teorema devido a Serrin (1969). Além disso, se a condição de Serrin não for satisfeita, há um resultado de não-existência. A partir disso foi perguntado se um resultado similar valeria no espaço Hiperbólico. A finalidade desta tese é dar uma resposta afirmativa a esta pergunta, exibindo uma condição tipo Serrin. De maneira que obtém-se existência de superfícies cujo gráfico tenha curvatura média hiperbólica pré-determinada H(x) no espaço hiperbólico. O resultado é sharp no sentido que se tal condição for negada então não-existência pode ser estabelecida. O ponto central é uma estimativa a priori do gradiente de uma tal solução. / [en] A classical result in Partial Differential Equations and Differential Geometrydue to Serrin (1969) is the following: Given a constant (alfa) there exists a condition on the boundary of the domain (omega)such that the Dirichlet problem for the mean equation (alfa)is solvable. Besides, if Serrin's condition fails there is a non-existence result. Taking into account this classical result one may ask if a similar theorem holds in hyperbolic space. The goal of this thesis is to give a positive answer to this question establishing a certain Serrin type condition. Thus we obtain existence of surfaces whose graphs has prescribed mean curvature H(x) in hyperbolic space. This result is sharp because if the condition is not satisfied then a non- existence result can be inferred. The main point of the argument is some a priori gradient estimate and degree theory.

Neste trabalho mostramos que dado um domínio exterior de classe C0 contido em uma superfície umb lica de H3; com curvatura média constante H 2 [0; 1); existe uma família de gracos de Killing com curvatura média constante H: O bordo de cada um destes gracos está contido nesta superfície umbílica e a norma do gradiente da função no bordo pode ser prescrita por um certo valor s 0. / In this paper we show that given an exterior domain of class C0 contained in an umbilical surface of H3; with constant mean curvature H 2 [0; 1); there exists a family of Killing graphs with constant mean curvature H: The boundary of each of these graphs is contained in this umbilical surface and the norm of the gradient of the function in the boundary can be prescribed by a certain value s 0:

The investigation of the latent geometrical space behind complex network topologies is a fervid topic in current network science and the hyperbolic space is one of the most studied, because it seems associated to the structural organization of many real complex systems. The popularity-similarity-optimization (PSO) generative model is able to grow random geometric graphs in the hyperbolic space with realistic properties such as clustering, small-worldness, scale-freeness and rich-clubness. However, it misses to reproduce an important feature of real complex systems, which is the community organization. Here, we introduce the nonuniform PSO (nPSO) generative model, a generalization of the PSO model with a tailored community structure, and we provide an efficient algorithmic implementation with a O(EN) time complexity, where N is the number of nodes and E the number of edges. Meanwhile, in recent years, the inverse problem has also gained increasing attention: given a network topology, how to provide an accurate mapping into its latent geometrical space. Unlike previous attempts based on a computationally expensive maximum likelihood optimization (whose time complexity is between O(N^3) and O(N^4)), here we show that a class of methods based on nonlinear dimensionality reduction can solve the problem with higher precision and reducing the time complexity to O(N^2).

Knowledge graphs (KGs) store relational information in a flexible triplet schema and have become ubiquitous for information storage in domains such as web search, e-commerce, social networks, and biology. Retrieval of information from KGs is generally achieved through logical reasoning, but this process can be computationally expensive and has limited performance due to the large size and complexity of relationships within the KGs. Furthermore, to extend the usage of KGs to non-expert users, retrieval over them cannot solely rely on logical reasoning but also needs to consider text-based search. This creates a need for multi-modal representations that capture both the semantic and structural features from the KGs. The primary objective of the proposed work is to extend the accessibility of KGs to non-expert users/institutions by enabling them to utilize non-technical textual queries to search over the vast amount of information stored in KGs. To achieve this objective, the research aims to solve four limitations: (i) develop a framework for logical reasoning over KGs that can learn representations to capture hierarchical dependencies between entities, (ii) design an architecture that can effectively learn the logic flow of queries from natural language text, (iii) create a multi-modal architecture that can capture inherent semantic and structural features from the entities and KGs, respectively, and (iv) introduce a novel hyperbolic learning framework to enable the scalability of hyperbolic neural networks over large graphs using meta-learning. The proposed work is distinct from current research because it models the logical flow of textual queries in hyperbolic space and uses it to perform complex reasoning over large KGs. The models developed in this work are evaluated on both the standard research setting of logical reasoning, as well as, real-world scenarios of query matching and search, specifically, in the e-commerce domain. In summary, the proposed work aims to extend the accessibility of KGs to non-expert users by enabling them to use non-technical textual queries to search vast amounts of information stored in KGs. To achieve this objective, the work proposes the use of multi-modal representations that capture both semantic and structural features from the KGs, and a novel hyperbolic learning framework to enable scalability of hyperbolic neural networks over large graphs. The work also models the logical flow of textual queries in hyperbolic space to perform complex reasoning over large KGs. The models developed in this work are evaluated on both the standard research setting of logical reasoning and real-world scenarios in the e-commerce domain. / Doctor of Philosophy / Knowledge graphs (KGs) are databases that store information in a way that allows computers to easily identify relationships between different pieces of data. They are widely used in domains such as web search, e-commerce, social networks, and biology. However, retrieving information from KGs can be computationally expensive, and relying solely on logical reasoning can limit their accessibility to non-expert users. This is where the proposed work comes in. The primary objective is to make KGs more accessible to non-experts by enabling them to use natural language queries to search the vast amounts of information stored in KGs. To achieve this objective, the research aims to address four limitations. Firstly, a framework for logical reasoning over KGs that can learn representations to capture hierarchical dependencies between entities is developed. Secondly, an architecture is designed that can effectively learn the logic flow of queries from natural language text. Thirdly, a multi-modal architecture is created that can capture inherent semantic and structural features from the entities and KGs, respectively. Finally, a novel hyperbolic learning framework is introduced to enable the scalability of hyperbolic neural networks over large graphs using meta-learning. The proposed work is unique because it models the logical flow of textual queries in hyperbolic space and uses it to perform complex reasoning over large KGs. The models developed in this work are evaluated on both the standard research setting of logical reasoning, as well as, real-world scenarios of query matching and search, specifically, in the e-commerce domain. In summary, the proposed work aims to make KGs more accessible to non-experts by enabling them to use natural language queries to search vast amounts of information stored in KGs. To achieve this objective, the work proposes the use of multi-modal representations that capture both semantic and structural features from the KGs, and a novel hyperbolic learning framework to enable scalability of hyperbolic neural networks over large graphs. The work also models the logical flow of textual queries in hyperbolic space to perform complex reasoning over large KGs. The results of this work have significant implications for the field of information retrieval, as it provides a more efficient and accessible way to retrieve information from KGs. Additionally, the multi-modal approach taken in this work has potential applications in other areas of machine learning, such as image recognition and natural language processing. The work also contributes to the development of hyperbolic geometry as a tool for modeling complex networks, which has implications for fields such as network science and social network analysis. Overall, this work represents an important step towards making the vast amounts of information stored in KGs more accessible and useful to a wider audience.

[pt] Seja S uma superfície fechada orientável, de gênero > 2 e sem pontos conjugados. Seja F uma folheação no fibrado tangente unitário de S, de codimensão 1, invariante pelo fluxo geodésico e de classe C2. Então, a curvatura de S é constante < 0. A demonstração é conseqüência dos dois seguintes resultados, que têm interesse por si mesmos. O primeiro é que se T1S admite uma folheação contínua de codimensão 1 por folhas C1 invariantes pelo fluxo geodésico então a superfície não tem pontos conjugados e a folheação coincide com a folheação centro-estável ou com a centro-instável. O segundo resultado é o seguinte. Seja S uma superfície fechada orientável, de gênero > 2 e sem pontos conjugados. Então, a folheação centro-estável Fcs de T1S é conjugada à folheação centro-estável da métrica hiperbólica em S. Esta conjugação é da mesma classe de diferenciabilidade de Fcs . Portanto, se Fcs é de classe C2, uma extensão da teoria de Godbillon-Vey implica que a curvatura da superfície é constante negativa. / [en] Lets be a orientable closed surface with no conjugate points. Let F be a foliation in the unitary tangent fiber bundle of S, of codimension 1, invariant by the geodesic flow and of class C2. Then, the curvature of S is constant < 0 . The demonstration is a consequence of the two following results, which are of interest by themselves. The first one is that if T1S admits a continuous foliation of codimension 1 by leaves C1 invariants by the geodesic flow, then the surface is with no conjugate points, and the foliation coincides with either the center stable foliation or the center unstable foliation. The second result is the following. Let S be a orientable closed surface of genus > 2 and with no conjugate points. Then, the center unstable foliation Fcs of T1S is conjugate to the center stable foliation of the hyperbolic metric in S. This conjugation is of the same class of differentiability of Fcs. Therefore, if Fcs is of class C2, an extension of the Godbillon-Vey theory implies that the curvature of the surface is constant negative.