Advances in additive manufacturing technologies have brought a new paradigm shift to both design and manufacturing. There is a much bigger design space in which designers can achieve a level of complexity and customizability, which are infeasible using traditional manufacturing processes. One application of this technology is for fabrication of meso-scale lattice structures (MSLS). These types of structures are designed to have material where it is needed for specific applications. They are suitable for any weight-critical applications, particularly in industries where both low weight and high strength are desired. MSLS can easily have hundreds to thousands of individual strut, where the diameter of each strut can be treated as a design variable. As a result, the design process poses a computational challenge. Since the computational complexity of the design problem often scales exponentially with the number of design variables, topological optimization that requires multi-variable optimization algorithm is infeasible for large-scale problems. In previous research, a new method was presented for efficiently optimizing MSLS by utilizing a heuristic that reduces the multivariable optimization problem to a problem of only two variables. The method is called the Size Matching and Scaling (SMS) method, which combines solid-body analysis and predefined unit-cell library to generate the topology of the structure. However, the method lacks a systematic methodology to generate the initial ground geometry for the design process, which limits the previous implementations of the SMS method to only simple, axis-aligned structures. In this research, an augmented SMS method is presented. The augmented method includes the integration of free-mesh approach in generating the initial ground geometry. The software that embodies that ground geometry generation process is integrated to commercial CAD system that allows designer to set lattice size parameters through graphical user interface. In this thesis, the augmented method and the unit-cell library are applied to various design examples. The augmented SMS method can be applied effectively in the design of conformal lattice structure with highly optimized stiffness and volume for complex surface. Conformal lattice structures are those conformed to the shape of a part's surface and that can used to stiffen or strengthen a complex and curved surface. This design approach removes the need for a rigorous topology optimization, which is a main bottleneck in designing MSLS.

Thesis (M.S.)--State University of New York at Binghamton, Department of Computer Science, Thomas J. Watson School of Engineering and Applied Science, 2007. / Includes bibliographical references.

Thesis (Ph.D.)--University of Cincinnati, 2008. / Committee/Advisors: David Herron PhD (Committee Chair), David Minda PhD (Committee Member), Nageswari Shanmugalingam PhD (Committee Member). Includes bibliographical references and abstract.

Thesis (Ph. D.)--University of Chicago, Department of Mathematics, June 2000. / Includes bibliographical references. Also available on the Internet.

Thesis (Ph. D.)--University of Washington, 2008. / Vita. Includes bibliographical references (p. 113-122).

Thesis (Ph. D.)--University of Washington, 2006. / Vita. Includes bibliographical references (p. 81-85).

abstract: The fields of pattern recognition and machine learning are on a fundamental quest to design systems that can learn the way humans do. One important aspect of human intelligence that has so far not been given sufficient attention is the capability of humans to express when they are certain about a decision, or when they are not. Machine learning techniques today are not yet fully equipped to be trusted with this critical task. This work seeks to address this fundamental knowledge gap. Existing approaches that provide a measure of confidence on a prediction such as learning algorithms based on the Bayesian theory or the Probably Approximately Correct theory require strong assumptions or often produce results that are not practical or reliable. The recently developed Conformal Predictions (CP) framework - which is based on the principles of hypothesis testing, transductive inference and algorithmic randomness - provides a game-theoretic approach to the estimation of confidence with several desirable properties such as online calibration and generalizability to all classification and regression methods. This dissertation builds on the CP theory to compute reliable confidence measures that aid decision-making in real-world problems through: (i) Development of a methodology for learning a kernel function (or distance metric) for optimal and accurate conformal predictors; (ii) Validation of the calibration properties of the CP framework when applied to multi-classifier (or multi-regressor) fusion; and (iii) Development of a methodology to extend the CP framework to continuous learning, by using the framework for online active learning. These contributions are validated on four real-world problems from the domains of healthcare and assistive technologies: two classification-based applications (risk prediction in cardiac decision support and multimodal person recognition), and two regression-based applications (head pose estimation and saliency prediction in images). The results obtained show that: (i) multiple kernel learning can effectively increase efficiency in the CP framework; (ii) quantile p-value combination methods provide a viable solution for fusion in the CP framework; and (iii) eigendecomposition of p-value difference matrices can serve as effective measures for online active learning; demonstrating promise and potential in using these contributions in multimedia pattern recognition problems in real-world settings. / Dissertation/Thesis / Ph.D. Computer Science 2010

The use of conformal transformation as a method for generating solutions of Einstein's equations has been mainly studied in the cases where the original spacetime is vacuum. The generated spacetimes then frequently belong to the class of pp-waves. In the present work, the electrovacuum spacetimes are stud- ied, i.e the solutions of coupled Einstein's and Maxwell's equations. By using the conformal transformation, it is possible to circumvent solving the later equa- tions. This method is concretely studied for null Einstein-Maxwell fields and it turns out that the admissible spacetimes are pp-waves again. However, if the method is generalized, it is possible to enlarge the class of conformal null Einstein-Maxwell fields to a wider family of Kundt spacetimes. 1

In this thesis we study the time evolution of Rényi and entanglement entropies of thermal states in Conformal Field Theory (CFT). These quantities are usually hard to compute but Ryu-Takayanagi (RT) and Hubeny-Rangamani-Takayanagi (HRT) proposals allow us to find the same quantities using calculations in general relativity. We will introduce main concepts of holography, quantum information and conformal field theory that will be used to derive the results of this thesis. In the first part of the thesis, we explicitly compute entanglement entropy of the rotating BTZ black hole by directly applying HRT proposal and finding lengths of spacelike geodesics. Rényi entropy of thermal state perturbed by a local quantum quench is computed by mapping correlators on two glued cylinders to the plane for field theory containing a single free boson and for 2d CFTs in the large c limit. We consider Thermofield Double State (TFD) which is an entangled state in direct product of two 2D CFTs. It is conjectured to be holographically equivalent to the eternal BTZ black hole. TFD state is perturbed by a local quench in one CFT and mutual information between two intervals in two CFTs is computed. We find when mutual information vanishes and interpret this as scrambling time, i.e. time scale required for the system to thermalize. This field theory result is modelled with a massive free falling particle in the BTZ black hole. We have computed the back-reaction of the particle on the metric of BTZ and used RT proposal to find holographic entanglement entropy. Finally, we generalize this calculation to the case of rotating BTZ with inner and outer horizons. It is dual to the CFT with different temperatures for left and right moving modes. We calculate mutual information and scrambling time and find exact agreement between results in the gravity and those in the CFT.

Static, spherically-symmetric particle-like solutions to the coupled Einstein-Dirac and Einstein-Dirac-Maxwell equations have been studied by Finster, Smoller and Yau (FSY). In their work, FSY left the fermion mass as a parameter set to ±1. This thesis generalises these equations to include the Higgs field, letting the fermion mass become a function through coupling, μ. We discuss the dynamics associated with the Higgs field and find that there exist qualitatively similar solutions to those found by FSY, with well behaved, non-divergent metric components and electrostatic potential, close to the origin, going over to the point-particle solutions for large r; the Schwarzschild or Reissner-Nordström metric, and the Coulomb potential. We then go on to discuss an alternative gravity theory, conformal gravity, (CG), and look for solutions of the CG equations of motion coupled to the Dirac, Higgs and Maxwell equations. We obtain asymptotically nonvanishing, yet fully normalisable Dirac spinor components, resembling those of FSY, and, in the case where charge is included, non-divergent electrostatic potential close to the origin, matching onto the Coulomb potential for large r.