Applied Mathematics Commons^™

Multi-Grade Deep Learning, Yuesheng Xu

Mathematics & Statistics Faculty Publications

Deep learning requires solving a nonconvex optimization problem of a large size to learn a deep neural network (DNN). The current deep learning model is of a single-grade, that is, it trains a DNN end-to-end, by solving a single nonconvex optimization problem. When the layer number of the neural network is large, it is computationally challenging to carry out such a task efficiently. The complexity of the task comes from learning all weight matrices and bias vectors from one single nonconvex optimization problem of a large size. Inspired by the human education process which arranges learning in grades, we …

Suntan (And Other Solar Trigonometric Functions): Solutions For Fermi Questions: February 2025, John Adam

Mathematics & Statistics Faculty Publications

The article discusses the intensity of sunlight on the side of the author's face as they walk to and from their office, focusing on the solar zenith angle θ. It presents a formula for solar irradiance and explores how the intensity changes based on the angle and elevation. The solutions to the questions posed in the article provide insights into the maximum irradiance levels at different angles and elevations.

Comprehensive Benchmarking Of Several Machine Learning And Bayesian Models For Early-Stage Diabetes Risk Prediction: A Large-Scale Comparative Study, Md. Iqbal Hossain, Najila Alam Porno

Mathematics & Statistics Faculty Publications

Diabetes remains a critical global health challenge, with early detection is crucial for effective management. This study presents a comprehensive benchmarking analysis of 14 diverse machine learning and Bayesian models for early-stage diabetes risk prediction using clinical data [2] from Sylhet, Bangladesh. This research evaluated traditional methods (Logistic Regression, Decision Trees), ensemble techniques (Random Forest, XGBoost, LightGBM), Bayesian approaches (BART, Bayesian Logistic Regression), and advanced neural architectures (Deep Belief Networks) using both 70-30 train-test splits and 10-fold cross-validation. The results demonstrate that ensemble methods consistently outperformed other approaches, with Random Forest(RF) achieving the highest cross-validated AUC (0.9951) and accuracy (0.9699). …

Suntan (And Other Solar Tigonometric Functions), John Adam

Mathematics & Statistics Faculty Publications

Question 1: If I₀ is the solar irradiance (power per unit area, W/m²) reaching my head, express the intensity on the side of my face (I_s) in terms of θ. Assume for now that the irradiance is independent of path length through the atmosphere and that my face is normal to the direction θ = 90°.

Using the 1962 U.S. Standard Atmosphere,² Hottel (1976)³ expressed the solar irradiance using the formula

I = I₀(a₀ + a₁e^{−k sec θ}), where A is the elevation in kilometers and

a₀ = 0.4237 − 0.00821(6 − A)²; a₁ = 0.5055 …

Golden Spirals Everywhere?, John Adam

Mathematics & Statistics Faculty Publications

The article explores different types of spirals, including Archimedean, hyperbolic, and logarithmic spirals, with a focus on the golden ratio and golden spirals. It discusses the misconception that golden rectangles and spirals can be found in various natural and man-made objects, emphasizing the importance of understanding the properties of logarithmic spirals. The text provides mathematical equations for logarithmic spirals and poses questions for readers to explore the concept further. The author, John Adam, invites readers to engage in Fermi Questions and submit ideas for consideration.

Investigations Of Conjugate Heat Transfer And Fluid Flow In Partitioned Porous Cavity Using Darcy-Forchheimer Model: Finite Element-Based Computations, Nasir Yasin, Shafee Ahmad, Muhammad Umair, Zahir Shah, Narcisa Vrinceanu, Ghadah Alhawael

Mathematics & Statistics Faculty Publications

The conjugate heat transfer and fluid flow has vast applications in thermal engineering, particularly for cooling in thermal devices, and automobile engines. This study investigates conjugate heat transfer in 2D enclosures, featuring thin solid fins attached to a porous bottom wall. The porous medium is considered isotropic and homogeneous by the Darcy-Forchheimer model, with fluid phases in local thermal equilibrium. The boundary conditions at the porous fluid interface ensure continuity of the velocities, stresses, temperature, and heat flux. The phenomenon is mathematically modelled by obtaining a set of partial differential equations. The finite element method (FEM) is used to perform …

Golden Spirals Everywhere? Solutions For Fermi Questions, January 2025, John Adam

Mathematics & Statistics Faculty Publications

The article discusses different types of spirals, including Archimedean, hyperbolic, and logarithmic spirals, with a focus on the golden ratio and golden spirals. It addresses the misconception that golden rectangles and spirals can be found in various natural and man-made structures, emphasizing the importance of understanding the properties of logarithmic spirals. The article provides mathematical explanations and solutions for questions related to pitch angles and self-similarity in logarithmic spirals, using examples like the nautilus shell and an ammonite-like stone. It concludes by referencing additional sources for further exploration of the golden ratio and golden spiral myths.

A Time-Domain Boundary Integral Equation For Moving Acoustic Sources In Uniform Flow And Its Solution By An Advanced Time Propagation Approach, Fang Q. Hu, Douglas M. Nark

Mathematics & Statistics Faculty Publications

This paper presents a time-domain boundary integral equation (TDBIE) formulation for predicting acoustic scattering from moving sources in a uniform mean flow. This work is motivated by the increasing need for accurate aeroacoustic modeling of modern aircraft configurations, including VTOL and eVTOL systems with rotating components. A key challenge in time-domain scattering simulations with moving sources is the determination of retarded time for a given observer time, which involves solving an implicit equation at each time step. This can be computationally costly, particularly for numerical solution of the TDBIE where every surface element on the scattering body acts as an …

Covariate Selection For Rna-Seq Differential Expression Analysis With Hidden Factor Adjustment, Farzana Noorzahan, Hyeongseon Jeon, Yet Nguyen

Mathematics & Statistics Faculty Publications

In RNA-seq data analysis, a primary objective is the identification of differentially expressed genes, which are genes that exhibit varying expression levels across different conditions of interest. It is widely known that hidden factors, such as batch effects, can substantially influence the differential expression analysis. Furthermore, apart from the primary factor of interest and unforeseen artifacts, an RNA-seq experiment typically contains multiple measured covariates, some of which may significantly affect gene expression levels, while others may not. Existing methods either address the covariate selection or the unknown artifacts separately. In this study, we investigate two integrated strategies, FSR_sva and SVAall_FSR, …

Maximum Trimmed Likelihood Estimation For Discrete Multivariate Vasicek Processes, Thomas M. Fullerton Jr., Michael Pokojovy, Andrews T. Anum, Ebenezer Nkum

Mathematics & Statistics Faculty Publications

The multivariate Vasicek model is commonly used to capture mean-reverting dynamics typical for short rates, asset price stochastic log-volatilities, etc. Reparametrizing the discretized problem as a VAR(1) model, the parameters are oftentimes estimated using the multivariate least squares (MLS) method, which can be susceptible to outliers. To account for potential model violations, a maximum trimmed likelihood estimation (MTLE) approach is utilized to derive a system of nonlinear estimating equations, and an iterative procedure is developed to solve the latter. In addition to robustness, our new technique allows for reliable recovery of the long-term mean, unlike existing methodologies. A set of …

Weak-Strong Beam-Beam Simulation With Crab Cavity Noises For The Hadron Storage Ring Of The Electron-Ion Collider, Y. Luo, B. Gamage, C. Montag, D. Marx, D. Xu, F. Willeke, H. Huang, H. Lovelace Iii, J. Berg, M. Blaskiewicz, S. Peggs, T. Satogata, V. Ptitsyn, V. Morozov, Y. Hao

Mathematics & Statistics Faculty Publications

The Electron Ion Collider (EIC), to be constructed at Brookhaven National Laboratory, will collide polarized high-energy electron beams with hadron beams, achieving luminosities of up to 1 X 10³⁴cm⁻²s⁻¹ in the center-mass energy range of 20-140 GeV. Crab cavities are employed to compensate for the geometric luminosity loss caused by a large crossing angle of 25 mrad in the interaction region. The phase noise in crab cavities will induce a significant emittance growth for the hadron beams in the Hadron Storage Ring (HSR). Various models have been utilized to study the effects of crab cavity …

A Novel Computational Analysis Of Boundary-Driven Two-Dimensional Heat Flow With Internal Heat Generation, Muhammad Abid, Madiha Bibi, Nasir Yasin, Muhammad Shahid

Mathematics & Statistics Faculty Publications

Accurate numerical solution of parabolic and elliptic partial differential equations governing two-dimensional heat transfer is critical for engineering simulations but computationally challenging. This work employs key numerical techniques finite differences, conjugate gradients, and Crank-Nicolson time stepping to solve the heat diffusion equation and analyze method performance. The Poisson equation is discretized using second-order central finite differences and solved with the conjugate gradient approach to determine the steady state solution. The transient heat equation is integrated in time via the Crank-Nicolson implicit scheme, also utilizing conjugate gradients. The methods effectively compute solutions matching analytical and boundary conditions. Convergence and stability are …

Uniform Convergence Of Deep Neural Networks With Lipschitz Continuous Activation Functions And Variable Widths, Yuesheng Xu, Haizhang Zhang

Mathematics & Statistics Faculty Publications

We consider deep neural networks (DNNs) with a Lipschitz continuous activation function and with weight matrices of variable widths. We establish a uniform convergence analysis framework in which sufficient conditions on weight matrices and bias vectors together with the Lipschitz constant are provided to ensure uniform convergence of DNNs to a meaningful function as the number of their layers tends to infinity. In the framework, special results on uniform convergence of DNNs with a fixed width, bounded widths and unbounded widths are presented. In particular, as convolutional neural networks are special DNNs with weight matrices of increasing widths, we put …

Automatic Hemorrhage Segmentation In Brain Ct Scans Using Curriculum-Based Semi-Supervised Learning, Solayman H. Emon, Tzu-Liang (Bill) Tseng, Michael Pokojovy, Peter Mccaffrey, Scott Moen, Md Fashiar Rahman

Mathematics & Statistics Faculty Publications

One of the major neuropathological consequences of traumatic brain injury (TBI) is intracranial hemorrhage (ICH), which requires swift diagnosis to avert perilous outcomes. We present a new automatic hemorrhage segmentation technique via curriculum-based semi-supervised learning. It employs a pre-trained lightweight encoder-decoder framework (MobileNetV2) on labeled and unlabeled data. The model integrates consistency regularization for improved generalization, offering steady predictions from original and augmented versions of unlabeled data. The training procedure employs curriculum learning to progressively train the model at diverse complexity levels. We utilize the PhysioNet dataset to train and evaluate the proposed approach. The performance results surpass those of …

Addressing Spectral Bias Of Deep Neural Networks By Multi-Grade Deep Learning, Ronglong Fang, Yuesheng Xu

Mathematics & Statistics Faculty Publications

Deep neural networks (DNNs) have showcased their remarkable precision in approximating smooth functions. However, they suffer from the spectral bias, wherein DNNs typically exhibit a tendency to prioritize the learning of lower-frequency components of a function, struggling to effectively capture its high-frequency features. This paper is to address this issue. Notice that a function having only low frequency components may be well-represented by a shallow neural network (SNN), a network having only a few layers. By observing that composition of low frequency functions can effectively approximate a high-frequency function, we propose to learn a function containing high-frequency components by composing …

Modeling The Spread Of Covid-19 In Spatio-Temporal Context, S.H. Sathish Indika, Norou Diawara, Hueiwang Anna Jeng, Bridget D. Giles, Dilini S.K. Gamage

Mathematics & Statistics Faculty Publications

This study aims to use data provided by the Virginia Department of Public Health to illustrate the changes in trends of the total cases in COVID-19 since they were first recorded in the state. Each of the 93 counties in the state has its COVID-19 dashboard to help inform decision makers and the public of spatial and temporal counts of total cases. Our analysis shows the differences in the relative spread between the counties and compares the evolution in time using Bayesian conditional autoregressive framework. The models are built under the Markov Chain Monte Carlo method and Moran spatial correlations. …

Not Your Typical Tower Of Sauron, John Adam

Mathematics & Statistics Faculty Publications

The picture is of the tapering Chester Shot Tower, located in Chester, England. It was built in 1799 for the manufacture of lead shot for use in the Napoleonic Wars. Molten lead was poured through a sieve at the top of the tower, with the tiny droplets forming perfect spheres during the fall; these were then cooled in a vat of water at the base. This process was less labor intensive than an earlier method using molds. It is the oldest of the three remaining shot towers in the UK.

Question 1: Using the parked van at the base, estimate …

Drop In The Bucket?, John Adam

Mathematics & Statistics Faculty Publications

No abstract provided.

Horizontal Air Mass: Solutions For Fermi Questions, November 2022, John Adam

Mathematics & Statistics Faculty Publications

No abstract provided.

Multiplicative Noise Removal: Nonlocal Low-Rank Model And It's Proximal Alternating Reweighted Minimization Algorithm, Xiaoxia Liu, Jian Lu, Lixin Shen, Chen Xu, Yuesheng Xu

Mathematics & Statistics Faculty Publications

The goal of this paper is to develop a novel numerical method for efficient multiplicative noise removal. The nonlocal self-similarity of natural images implies that the matrices formed by their nonlocal similar patches are low-rank. By exploiting this low-rank prior with application to multiplicative noise removal, we propose a nonlocal low-rank model for this task and develop a proximal alternating reweighted minimization (PARM) algorithm to solve the optimization problem resulting from the model. Specifically, we utilize a generalized nonconvex surrogate of the rank function to regularize the patch matrices and develop a new nonlocal low-rank model, which is a nonconvex …

Mesoscopic Methods In Engineering And Science, Christian Jansen, Manfred Krafczyk, Li-Shi Luo

Mathematics & Statistics Faculty Publications

(First paragraph) Matter, conceptually classified into fluids and solids, can be completely described by the microscopic physics of its constituent atoms or molecules. However, for most engineering applications a macroscopic or continuum description has usually been sufficient, because of the large disparity between the spatial and temporal scales relevant to these applications and the scales of the underlying molecular dynamics. In this case, the microscopic physics merely determines material properties such as the viscosity of a fluid or the elastic constants of a solid. These material properties cannot be derived within the macroscopic framework, but the qualitative nature of the …

Boundary Vortex Formation In Polarization-Modulated Orthogonal Smectic Liquid Crystals, Carlos J. García-Cervera, Tiziana Giorgi, Sookyung Joo

Mathematics & Statistics Faculty Publications

We investigate the relaxation of an energy functional originated in the physics literature to study the bistability of polarization modulated orthogonal smectic phases (SmAP_Fmod) of bent-core molecules liquid crystals. We show that the interplay between the mixed boundary conditions and the shape of the sample results in boundary defects. We also analyze the bistable switching due to an applied electric field via gradient flow numerical simulations. Our computations reveal a novel dynamic scenario, where switching is achieved by the formation of two internal vortices.

Dynamic Attribute-Level Best Worst Discrete Choice Experiments, Amanda Working, Mohammed Alqawba, Norou Diawara

Mathematics & Statistics Faculty Publications

Dynamic modelling of decision maker choice behavior of best and worst in discrete choice experiments (DCEs) has numerous applications. Such models are proposed under utility function of decision maker and are used in many areas including social sciences, health economics, transportation research, and health systems research. After reviewing references on the study of such experiments, we present example in DCE with emphasis on time dependent best-worst choice and discrimination between choice attributes. Numerical examples of the dynamic DCEs are simulated, and the associated expected utilities over time of the choice models are derived using Markov decision processes. The estimates are …

Numerical Simulation For A Rising Bubble Interacting With A Solid Wall: Impact, Bounce, And Thin Film Dynamics, Changjuan Zhang, Jie Li, Li-Shi Luo, Tiezheng Qian

Mathematics & Statistics Faculty Publications

Using an arbitrary Lagrangian-Eulerian method on an adaptive moving unstructured mesh, we carry out numerical simulations for a rising bubble interacting with a solid wall. Driven by the buoyancy force, the axisymmetric bubble rises in a viscous liquid toward a horizontal wall, with impact on and possible bounce from the wall. First, our simulation is quantitatively validated through a detailed comparison between numerical results and experimental data. We then investigate the bubble dynamics which exhibits four different behaviors depending on the competition among the inertial, viscous, gravitational, and capillary forces. A phase diagram for bubble dynamics has been produced using …

The Use Of Item Response Theory In Survey Methodology: Application In Seat Belt Data, Mark K. Ledbetter, Norou Diawara, Bryan E. Porter

Mathematics & Statistics Faculty Publications

Problem: Several approaches to analyze survey data have been proposed in the literature. One method that is not popular in survey research methodology is the use of item response theory (IRT). Since accurate methods to make prediction behaviors are based upon observed data, the design model must overcome computation challenges, but also consideration towards calibration and proficiency estimation. The IRT model deems to be offered those latter options. We review that model and apply it to an observational survey data. We then compare the findings with the more popular weighted logistic regression. Method: Apply IRT model to the observed data …

Switching Mechanism In The B-1revtilted Phase Of Bent-Core Liquid Crystals, Carlos J. García-Cervera, Tiziana Giorgi, Sookyung Joo, Xin Yang Lu

Mathematics & Statistics Faculty Publications

The B_1RevTilted is a uniformly smectic tilted columnar phase in which the macroscopic polarization can be reorientated via electric field. To study the effects on the reorientation mechanism of the various physical parameters, we analyze a local, and a non-local Landau-de Gennes-type energy functional. For the case of large columnar samples, we show that both energies give the same qualitative behavior, with a relevant role played by the terms that describe the interaction between polarization and nematic directors. We also obtain existence of the L²-gradient flow in metric spaces for the full local energy.

Fire, Ice, Water, And Dirt: A Simple Climate Model, John Kroll

Mathematics & Statistics Faculty Publications

A simple paleoclimate model was developed as a modeling exercise. The model is a lumped parameter system consisting of an ocean (water), land (dirt), glacier, and sea ice (ice) and driven by the sun (fire). In comparison with other such models, its uniqueness lies in its relative simplicity yet yielding good results. For nominal values of parameters, the system is very sensitive to small changes in the parameters, yielding equilibrium, steady oscillations, and catastrophes such as freezing or boiling oceans. However, stable solutions can be found, especially naturally oscillating solutions. For nominally realistic conditions, natural periods of order 100kyrs are …

On A Time Domain Boundary Integral Equation Formulation For Acoustic Scattering By Rigid Bodies In Uniform Mean Flow, Fang Q. Hu, Michelle E. Pizzo, Douglas M. Nark

Mathematics & Statistics Faculty Publications

It has been well-known that under the assumption of a uniform mean flow, the acoustic wave propagation equation can be formulated as a boundary integral equation. However, the constant mean flow assumption, while convenient for formulating the integral equation, does not satisfy the solid wall boundary condition wherever the body surface is not aligned with the assumed uniform flow. A customary boundary condition for rigid surfaces is that the normal acoustic velocity be zero. In this paper, a careful study of the acoustic energy conservation equation is presented that shows such a boundary condition would in fact lead to source …

A Wasserstein Gradient Flow Approach To Poisson-Nernst-Planck Equations, David Kinderlehrer, Leinard Monsaingeon, Xiang Xu

Mathematics & Statistics Faculty Publications

The Poisson-Nernst-Planck system of equations used to model ionic transport is interpreted as a gradient flow for the Wasserstein distance and a free energy in the space of probability measures with finite second moment. A variational scheme is then set up and is the starting point of the construction of global weak solutions in a unified framework for the cases of both linear and non-linear diffusion. The proof of the main results relies on the derivation of additional estimates based on the flow interchange technique developed by Matthes et al. in [D. Matthes, R.J. McCann and G. Savare, Commun. Partial …

Multiplicative Noise Removal With A Sparsity-Aware Optimization Model, Jian Lu, Lixin Shen, Chen Xu, Yuesheng Xu

Mathematics & Statistics Faculty Publications

Restoration of images contaminated by multiplicative noise (also known as speckle noise) is a key issue in coherent image processing. Notice that images under consideration are often highly compressible in certain suitably chosen transform domains. By exploring this intrinsic feature embedded in images, this paper introduces a variational restoration model for multiplicative noise reduction that consists of a term reflecting the observed image and multiplicative noise, a quadratic term measuring the closeness of the underlying image in a transform domain to a sparse vector, and a sparse regularizer for removing multiplicative noise. Being different from popular existing models which focus …