Browsing by Author "Arnold, Rachel Florence"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Complex Analysis on Planar Cell Complexes
    Arnold, Rachel Florence (Virginia Tech, 2008-04-29)
    This paper is an examination of the theory of discrete complex analysis that arises from the framework of a planar cell complex. Construction of this theory is largely integration-based. A combination of two cell complexes, the double and its associated diamond complex, allows for the development of a discrete Cauchy Integral Formula.
  • Thumbnail Image
    The Discrete Hodge Star Operator and Poincaré Duality
    Arnold, Rachel Florence (Virginia Tech, 2012-05-01)
    This dissertation is a uniïfication of an analysis-based approach and the traditional topological-based approach to Poincaré duality. We examine the role of the discrete Hodge star operator in proving and in realizing the Poincaré duality isomorphism (between cohomology and homology in complementary degrees) in a cellular setting without reference to a dual cell complex. More specifically, we provide a proof of this version of Poincaré duality over R via the simplicial discrete Hodge star defined by Scott Wilson in [19] without referencing a dual cell complex. We also express the Poincaré duality isomorphism over both R and Z in terms of this discrete operator. Much of this work is dedicated to extending these results to a cubical setting, via the introduction of a cubical version of Whitney forms. A cubical setting provides a place for Robin Forman's complex of nontraditional differential forms, defined in [7], in the uniïfication of analytic and topological perspectives discussed in this dissertation. In particular, we establish a ring isomorphism (on the cohomology level) between Forman's complex of differential forms with his exterior derivative and product and a complex of cubical cochains with the discrete coboundary operator and the standard cubical cup product.
  • Thumbnail Image
    Learning Hyperparameters for Inverse Problems by Deep Neural Networks
    McDonald, Ashlyn Grace (Virginia Tech, 2023-05-08)
    Inverse problems arise in a wide variety of applications including biomedicine, environmental sciences, astronomy, and more. Computing reliable solutions to these problems requires the inclusion of prior knowledge in a process that is often referred to as regularization. Most regularization techniques require suitable choices of regularization parameters. In this work, we will describe new approaches that use deep neural networks (DNN) to estimate these regularization parameters. We will train multiple networks to approximate mappings from observation data to individual regularization parameters in a supervised learning approach. Once the networks are trained, we can efficiently compute regularization parameters for newly-obtained data by forward propagation through the DNNs. The network-obtained regularization parameters can be computed more efficiently and may even lead to more accurate solutions compared to existing regularization parameter selection methods. Numerical results for tomography demonstrate the potential benefits of using DNNs to learn regularization parameters.
  • Thumbnail Image
    The Role of Students' Gestures in Offloading Cognitive Demands on Working Memory in Proving Activities
    Kokushkin, Vladislav (Virginia Tech, 2023-02-03)
    This study examines how undergraduate students use hand gestures to offload cognitive demands on their working memory (WM) when they are engaged in three major proving activities: reading, presenting, and constructing proofs of mathematical conjectures. Existing research literature on the role of gesturing in cognitive offloading has been limited to the context of elementary mathematics but has shown promise for extension to the college level. My framework weaves together theoretical constructs from mathematics education and cognitive psychology: gestures, WM, and mathematical proofs. Piagetian and embodied perspectives allow for the integration of these constructs through positioning bodily activity at the core of human cognition. This framework is operationalized through the methodology for measuring cognitive demands of proofs, which is used to identify the set of mental schemes that are activated simultaneously, as well as the places of potential cognitive overload. The data examined in this dissertation includes individual clinical interviews with six undergraduate students enrolled in different sections of the Introduction to Proofs course in Fall 2021 and Spring 2022. Each student participated in seven interviews: two WM assessments, three proofs-based interviews, a stimulated recall interview (SRI), and post-interview assessments. In total, 42 interviews were conducted. The participants' hand gesturing and mathematical reasoning were qualitatively analyzed. Ultimately, students' reflections during SRIs helped me triangulate the initial data findings. The findings suggest that, in absence of other forms of offloading, hand gesturing may become a convenient, powerful, although not an exclusive offloading mechanism: several participants employed alternative mental strategies in overcoming the cognitive overload they experienced. To better understand what constitutes the essence of cognitive offloading via hand gesturing, I propose a typology of offloading gestures. This typology differs from the existing classification schemes by capturing the cognitive nuances of hand gestures rather than reflecting their mechanical characteristics or the underlying mathematical content. Employing the emerged typology, I then show that cognitive offloading takes different forms when students read or construct proofs, and when they present proofs to the interviewer. Finally, I report on some WM-related issues in presenting and constructing proofs that can be attributed to the potential side effects of mathematical chunking. The dissertation concludes with a discussion of the limitations and practical implications of this project, as well as foreshadowing the avenues for future research.