Abstract

This dissertation investigates the inherent instability of deep learning models, specifically focusing on the origin and mitigation of adversarial examples – imperceptible input changes that lead to catastrophic prediction errors. While neural networks (NNs) have achieved state- of-the-art results in fields ranging from image classification to hazard forecasting, their lack of reliability in safety-critical applications remain a significant barrier to their deployment as alternatives to traditional, time-intensive supercomputer simulations. The core contribu- tion of this work is the application of the low-rank Householder expansion (LRHE). The expansion provides a tractable way to identify unstable singular vectors through singular vector decomposition (SVD), which represents the specific directions in input space that trigger erroneous output changes. Furthermore, the dissertation introduces a novel way to expand max pooling layers using Householder reflectors, extending the LRHE framework to more complex architectures. Finally, a new adversarial training procedure based on LRHE regularization is proposed. This research aims to enhance the reliability of neural network models and applies the adversarial training to safety-critical applications such as tsunami early warning.

Committee Chair

Donsub Rim

Committee Members

Douglas Wiens; Mladen Wickerhauser; Rachel Roberts; Renato Feres

Degree

Doctor of Philosophy (PhD)

Author's Department

Mathematics

Author's School

Graduate School of Arts and Sciences

Document Type

Dissertation

Date of Award

8-3-2026

Language

English (en)

Available for download on Sunday, January 31, 2027

Included in

Mathematics Commons

Share

COinS