arXiv subjects

Dwight Nwaigwe

Publications and source records attributed to Dwight Nwaigwe.

6 recordsLinked to original sources

arXiv2024

On the existence of the maximum likelihood estimate and convergence rate under gradient descent for multi-class logistic regression

We revisit the problem of the existence of the maximum likelihood estimate for multi-class logistic regression. We show that one method of ensuring its existence is by assigning positive probability to every class in the sample dataset. The notion of data separability is not needed, which is in contrast to the classical set up of multi-class logistic regression in which each data sample belongs to one class. We also provide a general and constructive estimate of the convergence rate to the maximum likelihood estimate when gradient descent is used as the optimizer. Our estimate involves bounding the condition number of the Hessian of the maximum likelihood function. The approaches used in this article rely on a simple operator-theoretic framework.

cs.LG↗

arXiv2023

Graph-based methods coupled with specific distributional distances for adversarial attack detection

Artificial neural networks are prone to being fooled by carefully perturbed inputs which cause an egregious misclassification. These \textit{adversarial} attacks have been the focus of extensive research. Likewise, there has been an abundance of research in ways to detect and defend against them. We introduce a novel approach of detection and interpretation of adversarial attacks from a graph perspective. For an input image, we compute an associated sparse graph using the layer-wise relevance propagation algorithm \cite{bach15}. Specifically, we only keep edges of the neural network with the highest relevance values. Three quantities are then computed from the graph which are then compared against those computed from the training set. The result of the comparison is a classification of the image as benign or adversarial. To make the comparison, two classification methods are introduced: 1) an explicit formula based on Wasserstein distance applied to the degree of node and 2) a logistic regression. Both classification methods produce strong results which lead us to believe that a graph-based interpretation of adversarial attacks is valuable.

cs.LG↗

arXiv2020

Determinant of a Sum of Certain Kronecker Products

We compute the determinant of $\sum_{n=1}^{N} \vec{A}^{(n)} \otimes \vec{B}^{(n)}$, where $\vec{A}^{(n)}$ is square and ${\vec{B}^{(n)}=\vec{x}^{(n)}{\vec{y}^{(n)}}^T}$ where $\vec{x}^{(n)}$ and $\vec{y}^{(n)}$ have length $N$.

math.CO↗

arXiv2020

On the Convergence of WKB Approximations of the Damped Mathieu Equation

Consider the differential equation ${ m\ddot{x} +γ\dot{x} -xε\cos(ωt) =0}$, $0 \leq t \leq T$. The form of the fundamental set of solutions are determined by Floquet theory. In the limit as $m \to 0$ we can apply WKB theory to get first order approximations of this fundamental set. WKB theory states that this approximation gets better as $m \to 0$ in the sense that the difference in sup norm is bounded as function of $m$ for a given $T$. However, convergence of the periodic parts and exponential parts are not addressed. We show that there is convergence to these components. The asymptotic error for the characteristic exponents are $O(m^2)$ and $O(m)$ for the periodic parts.

math.CA↗

arXiv2020

Development of a New Image-to-text Conversion System for Pashto, Farsi and Traditional Chinese

We report upon the results of a research and prototype building project \emph{Worldly~OCR} dedicated to developing new, more accurate image-to-text conversion software for several languages and writing systems. These include the cursive scripts Farsi and Pashto, and Latin cursive scripts. We also describe approaches geared towards Traditional Chinese, which is non-cursive, but features an extremely large character set of 65,000 characters. Our methodology is based on Machine Learning, especially Deep Learning, and Data Science, and is directed towards vast quantities of original documents, exceeding a billion pages. The target audience of this paper is a general audience with interest in Digital Humanities or in retrieval of accurate full-text and metadata from digital images.

cs.CV↗

arXiv2020

Overdamped dynamics of a Brownian particle levitated in a Paul trap

We study the dynamics of the center of mass of a Brownian particle levitated in a Paul trap. We focus on the overdamped regime in the context of levitodynamics, comparing theory with our numerical simulations and experimental data from a nanoparticle in a Paul trap. We provide an exact analytical solution to the stochastic equation of motion, expressions for the standard deviation of the motion, and thermalization times by using the WKB method under two different limits. Finally, we prove the power spectral density of the motion can be approximated by that of an Ornstein-Uhlenbeck process and use the found expression to calibrate the motion of a trapped particle.

cond-mat.stat-mech↗