Physical Sciences and Mathematics Commons^™

Meta-Clustering For Specialized Language Models: Enhancing Contextual Adaptation And Mitigating Hallucinations In Diverse Healthcare Environments, Joshit Mohanty, Vaishali Vaishali, Sandeep Kumar Nayak, Sumit Lahiri

Graduate Student Government Association Research Conference

Large Language Models (LLMs) have significantly advanced conversational AI by enabling dialogic information-seeking and task execution across diverse domains. However, their extensive parameters and broad domain scope lead to “data hallucinations.” These shortcomings are particularly evident in dynamic and diverse environments like India’s healthcare sector, where myriad languages, regional practices, and cultural nuances demand specialized, localized expertise rather than one-size-fits-all generalist models. This paper introduces a meta-clustering framework that integrates Distilled Language Models (DLMs) and Small/Specialized Language Models (SLMs) with meta-learning principles to address these limitations. By drawing on evidence from works such as MedHalu and Med-HALT, the framework seeks …

Tamos: Task-Aware Multi-Agent Orchestrator System, Joshit Mohanty, Sandeep Kumar Nayak, Sumit Lahiri

Graduate Student Government Association Research Conference

Large language models (LLMs) are increasingly at the core of multi-agent systems (MAS). However, the high resource demand, error propagation, and lack of adaptive evaluation mechanisms pose significant challenges in deploying these agentic solutions at scale. To address these concerns, this research proposes a Task-Aware Multi-Agent Orchestrator System designed to refine the agentic framework, categorizing tasks autonomously, assigning specialized evaluation datasets, and balancing token usage against functional effectiveness. This approach underscores robust data management, including AsyncHow, Mosaic AI, and Synthetic Preference Optimization (PO) corpora. Each dataset targets specific dimensions of agent performance, such as dynamic task decomposition and tool integration …

Graphtreemed: A Hybrid Graph-Tree Rag Architecture For Mission-Critical Medical Applications, Joshit Mohanty, Sandeep Kumar Nayak, Sumit Lahiri

Graduate Student Government Association Research Conference

Studies within engineering management indicate that decision-making is often based on the cognitive processing of grouped and pictographic information clusters entangled with high-level pattern recognition. Similarly, graph-based retrieval-augmented generation (RAG) architectures substantially improve diagnostic accuracy and interpretability, while tree-structured systems reduce critical misses through hierarchical reasoning. However, existing solutions often lack a unified framework that seamlessly integrates these two paradigms to address the multifaceted demands of mission-critical healthcare settings. This proposal introduces GraphTreeMed, a novel hybrid RAG architecture designed to harness the complementary strengths of graph-based and tree-based retrieval mechanisms, thereby advancing the safety and efficacy of clinical decision support …

Building Climate Resilience Through Food Recovery, San Diego Regional Climate Collaborative, Alexander Pourfard, Aranza Ruvalcaba

San Diego Regional Climate Collaborative

In an effort to combat GHG emissions from food waste, California passed Senate Bill 1383 (SB 1383) into law in 2016. The bill contains key provisions aimed at increasing edible food recovery. By rescuing food that would otherwise go to the landfill, California entities have been able to help save precious food and advance sustainable solutions to address climate change in their local communities. Since the food recovery requirements took effect in January of 2022, over 217,000 tons of unsold food have been recovered annually, providing more than 700 million meals to food-insecure Californians.

The Locus Of Intersecting Tangents, James Austin Canard

Student Theses and Dissertations

Consider two tangent lines to a given function, whose horizontal distance between the tangent points is a constant h. As the tangent lines change on the function, the intersection of the two tangent lines trace out a curve. We study the relationship between this new curve, called the locus curve, and the original function. We begin by studying the effects of vertical scaling, vertical shifting, vertical sheering, horizontal scaling, and horizontal shifting. Then we use asymptotical methods to study the behavior of the locus as the tangent lines approach infinity. We use a similar method to observe the behavior as …

Prime Factors Of The Third Cyclotomic Polynomial Of The Same Form And Inequalities Related To Odd Perfect Numbers, Cody Stenberg Hansen

Theses and Dissertations

This thesis relates some results about odd perfect numbers and factorizations of cyclotomic polynomials evaluated on integers. In Chapter 2 we parameterize solutions to the equality $\Phi_3(x)=\Phi_3(a_1)\Phi_3(a_2)\cdots\Phi_3(a_n)$, where $x,a_i\in \Z_{>0}$ for $1\leq i\leq n$, $\Phi_3$ is the third cyclotomic polynomial, and each $\Phi_3(a_i)$ is prime, in the cases that $n=2,3,4,5$. Let $N$ be an odd perfect number. Let $\omega(N)$ be the number of distinct prime factors of $N$ and let $\Omega(N)$ be the total number (counting multiplicity) of prime factors of $N$. In Chapter 3, we use a result from Chapter 2 to show that $\frac{99}{37}\omega(N) - \frac{187}{37} \leq …

Synthesis Of Novel Five-Membered Heterocycles As Bioactive Agents, Subrata Roy

Student Theses and Dissertations

This dissertation outlines the synthesis of various types of heterocyclic small molecules across three chapters. The first chapter discusses the efficient synthesis of seventy-one fluorinated pyrazole molecules and thirty-five 3,4-dihydroxy and 4-hydroxy phenyl-substituted thiazole molecules. These molecules were designed and synthesized for their potential antibacterial properties. The second chapter discusses the synthesis of fused thiazole molecules with enone-based steroidal natural products. This chapter reports on sixty-five novel fused thiazole molecules distributed across three libraries featuring different steroidal natural products: ethisterone, bisnoralcohol, and diosgenin. These molecules were designed and synthesized as potential anticancer agents. The third chapter comprises the development of …

A Novel Approach To Detect Shapes Of Constant Width, Ian D. Monk

Theses and Dissertations

Bodies of constant width have been studied mathematically since Euler and have found various applications in engineering and other areas. In this paper, we seek to develop novel tools to gain deeper insights into these shapes and to better be able to analyze them. The idea which motivates the development in this paper is as follows: a body has constant width if and only if the chord connecting the intersections of the boundary of the body with each pair of parallel hyper planes which lie tangent to a body lies perpendicular to the hyper planes.This is a known fact and …

04.21.2025 Ored Connect, Liz Williamson

ORED Newsletter

ORED Stakeholder Meetings

SEC Travel Grants

Innovative Machine Learning, Isotopic, And Hydrogeochemical Techniques For Groundwater Analysis In Arid Landscapes In Egypt’S Eastern Desert, Saad Ahmed Mohallel, Hesham Morgan, Ali Elgendy, Surendra Maharjan, Shahryar Fazli, Wenzhao Li, Rejoice Thomas, Hesham El-Askary

Mathematics, Physics, and Computer Science Faculty Articles and Research

Groundwater serves as a lifeline in Egypt’s hyper-arid Eastern Desert, particularly for agricultural and domestic uses. However, a comprehensive understanding of groundwater origin, quality, and recharge dynamics in the region remains limited due to geological complexity, data scarcity, and the high cost of isotopic analysis. This study addresses these challenges by integrating stable isotopes (δ¹⁸O and δ²H), hydrogeochemical parameters, remote sensing, and explainable artificial intelligence (AI) to investigate groundwater dynamics and support sustainable water management strategies. A total of 34 groundwater samples were collected from three key aquifers: the Quaternary alluvium, Nubian Sandstone, and fractured Basement aquifers. Hydrochemical analyses and …

Stability, Binding, And Charge Localization In Amorphous Tin Monoxide: A-Sno And A-Sn-Ta-O With Variable Composition, Julia E. Medvedeva, Joshua Santy, Mathew Pollard

Physics Faculty Research & Creative Works

The structural, electronic, and optical properties of amorphous SnO and Sn-Ta-O (with Sn:Ta ratios 5:1, 3:2, and 1:2) are investigated as a function of density and metal composition using ab initio molecular dynamics simulations and hybrid functional calculations. The short- and medium-range structures of amorphous oxides are thoroughly studied by calculating the effective coordination numbers, effective distances, distortion, and their time variances for metal-oxygen (M-O), O-M, M-M shells, the M-O-M and O-M-O angle distribution, as well as the volume, shape, and distribution of structural voids. Disorder is found to suppress both Sn-O and Sn-Sn coordination but not the O-Sn coordination. …

Stereotyping In Language (Technologies): An Examination Of Racial And Gender Stereotypes In Natural Language And Language Models, Messi Lee

McKelvey School of Engineering Graduate Student Theses & Dissertations

This dissertation examines stereotyping across natural language and language technologies through three interconnected studies. The first chapter applies a contemporary model of race relations from social psychology to investigate America's racial framework within American English, revealing how language encodes hierarchical associations between racial/ethnic groups and attributes of superiority and Americanness. The second chapter extends this analysis to Large Language Models (LLMs), finding that these language technologies portray socially subordinate groups as more homogeneous compared to dominant groups. The third chapter investigates stereotyping in Vision Language Models (VLMs), showing that these language technologies generate more uniform representations for women than men …

Trimming Five Generated Gorenstein Ideals, Luigi Ferraro, W. Frank Moore

School of Mathematical & Statistical Sciences Faculty Publications

Let (𝑅,𝔪,𝕜) be a regular local ring of dimension 3. Let I be a Gorenstein ideal of R of grade 3. It follows from a result of Buchsbaum and Eisenbud that there is a skew-symmetric matrix of odd size such that I is generated by the sub-maximal pfaffians of this matrix. Let J be the ideal obtained by multiplying some of the pfaffian generators of I by 𝔪; we say that J is a trimming of I. In a previous work, the first author and A. Hardesty constructed an explicit free resolution of 𝑅/𝐽 and computed a DG algebra …

Multiparticle Quantum Plasmonics: Fundamentals And Applications, Mingyuan Hong

LSU Doctoral Dissertations

Quantum plasmonics explores the interaction between light and collective charge oscillations at metal-dielectric interfaces, enabling strong light confinement and enhanced quantum effects at the nanoscale. While traditional quantum optics has primarily focused on single-photon systems, an intermediate regime exists between classical and single-photon optics - multiparticle (or multiphoton) quantum optics. In this regime, classical light sources, when analyzed through techniques such as photon-number-resolving (PNR) detection and projective measurement, can reveal nontrivial quantum correlations. This thesis investigates how multiparticle quantum plasmonics harnesses these correlations to control quantum statistical properties, enhance coherence, and enable novel applications in quantum technologies.

In this thesis, …

Impacts Of Extreme Weather Conditions On Coastal Fisheries Near Bayou Teche, Louisiana From 2019–2023, Georgia C. Davis

LSU Master's Theses

The region of Southcentral Louisiana, particularly around Bayou Teche, thrives on its commercial and recreational fisheries. These fisheries are occasionally subject to events like extreme weather that cause sudden and unexpected losses (NOAA Fisheries, 2024a). Between 2019 and 2023, a series of extreme weather events impacted southcentral Louisiana near Bayou Teche. This series of extreme events includes Hurricane Barry (2019), Hurricane Laura (2020), Hurricane Delta (2020), Hurricane Zeta (2020), Hurricane Ida (2021), and a United States Drought Monitor (USDM) D4 drought (2023). This study analyzes the impacts of the extreme weather series on coastal fish observed species richness (SR) in …

Shaping Mathematics Identity: An Exploratory Study On Specifications Grading In Calculus I At A Hispanic-Serving Institution, Luis M. Fernandez, Kaitlyn Stephens Serbin, Cristina Villalobos, Shaghayegh Azadi Setayesh, Guillermo Garza

School of Mathematical & Statistical Sciences Faculty Publications

Calculus I courses play a pivotal role in shaping students' STEM pathways, making it essential to adopt pedagogies that foster both achievement and mathematics identity development, particularly among underserved groups such as Hispanic students. This study explores the impact of Specifications Grading, an alternative assessment method where students meet specific course learning objectives through multiple attempts, on students’ mathematics identity development. Through a comparative case study of two Calculus I students at a Hispanic-Serving Institution, one enrolled in a specification graded course and the other in a traditionally-graded course, we examine shifts in their self-perceptions of competence, interest, recognition in …

2025 Academic Conference Program

Holy Cross Undergraduate Academic Conference

No abstract provided.

Rings Of Ice, Erik Johnson

Natural Science Faculty

Have you ever seen the Milky Way? For most Americans, it is impossible to see the band of light from the galaxy outside their homes. To be able to experience the night sky as our ancestors did, we have to travel to places away from light pollution. International Dark Sky Week, which is happening April 22–30 this year, celebrates the night sky and offers everyone a chance to reflect on the lighting choices made around their home and in their community.

Data-Driven Survival Modeling For Breast Cancer Prognostics: A Comparative Study With Machine Learning And Traditional Survival Modeling Methods, Theophilus Gyedu Baidoo, Hansapani Rodrigo

School of Mathematical & Statistical Sciences Faculty Publications

Background This investigation delves into the potential application of data-driven survival modeling approaches for prognostic assessments of breast cancer survival. The primary objective is to evaluate and compare the ability of machine learning (ML) models and conventional survival analysis techniques, to identify consistent key predictors of breast cancer survival outcomes.

Methods This study employs data-driven survival modeling approaches to predict breast cancer survival, including survival-specific methods such as the Cox Proportional Hazards (CPH) model, Random Survival Forests (RSF), and Cox Proportional Deep Neural Networks (DeepSurv), as well as machine learning models like Random Forests (RF), XGBoost, Support Vector Machines (SVM) …

Search For Continuous Gravitational Waves From Known Pulsars In The First Part Of The Fourth Ligo-Virgo-Kagra Observing Run, A. G. Abac, R. Abbott, I. Abouelfettouh, F. Acernese, Teviet Creighton, Mario C. Diaz, Francisco Llamas Villarreal, Soma Mukherjee, Volker Quetschke, Wenhui Wang

Physics & Astronomy Faculty Publications

Continuous gravitational waves (CWs) emission from neutron stars carries information about their internal structure and equation of state, and it can provide tests of general relativity. We present a search for CWs from a set of 45 known pulsars in the first part of the fourth LIGO–Virgo–KAGRA observing run, known as O4a. We conducted a targeted search for each pulsar using three independent analysis methods considering single-harmonic and dual-harmonic emission models. We find no evidence of a CW signal in O4a data for both models and set upper limits on the signal amplitude and on the ellipticity, which quantifies the …

Decadal To Millennial-Scale Morphodynamic Evolution Of Microtidal Marshes, Samuel Mason Zapp

LSU Doctoral Dissertations

Tidal marshes are susceptible to rapidly converting to open water due to their low elevations within the tidal frame. A continuous supply of externally supplied mineral sediment and in situ organic material production is necessary to offset elevation losses from relative sea level rise and erosion. Complex feedbacks between hydrodynamics, the erosion, transport, and deposition of sediment, and physical modification of the landscape by the establishment of intertidal vegetation govern the spatial and temporal patterns of elevation response. This dissertation focuses on modeling how these dynamics drive marsh landscape evolution in the past, present, and into the future under changing …

Predicting The Unpredictable: Predicting The March Madness Champion Using Statistical Modeling, Jack Sweeney

Honors Projects in Information Systems and Analytics

One of the more exciting and hardest parts of the men's college basketball postseason tournament, named March Madness, is to pick who wins each game and determine the overall champion of the tournament. The data analysis conducted will help determine the overall tournament winner. A machine learning model is implemented using college basketball metrics from previous years to determine the overall winner of this prestigious tournament in 2025. The model was able to pick up the winner in 2024. The result also finds that the most important features in determining the winner of the tournament are shooting guard height, small …

Analyzing The Declining Role Of Starting Pitchers In Mlb, Andrew Sferratore

Honors Projects in Information Systems and Analytics

In Major League Baseball (MLB) , the use of starting pitchers has faced major scrutiny. This study aims to discuss the declining role of starting pitchers in MLB and analyze how this affects team success during the regular season. Recent literature and statistics reveal that starting pitchers are throwing fewer innings and have lower pitch counts over the last two decades than in previous years. Attributing to this declining role are extreme spikes in the number of Tommy John (UCLR) surgeries, newly added rule changes, and the introduction of new philosophical approaches. Data from the 2024 regular season regarding starting …

Analyticity And Supershift With Regular Sampling, Fabrizio Colombo, Irene Sabadini, Daniele C. Struppa, Alain Yger

Mathematics, Physics, and Computer Science Faculty Articles and Research

The notion of supershift (in itself a generalization of the notion of superoscillation arising in quantum mechanics) expresses the fact that the sampling of a function in an interval allows to compute the values of the function far from the interval. In this paper, we study the relation between supershift and real analyticity. We use a classical result due to Serge Bernstein to show that real analyticity for a complex-valued function implies a strong form of supershift. On the other hand, we use a parametric version of a result by Leonid Kantorovitch to show that the converse is not true. …

Reinventing Mathematics Learning: An Exploration Of Undergraduate Mathematics Learning Post-Pandemic, Morgan R. Balesano

Honors Scholar Theses

In response to the COVID-19 pandemic, beginning in March 2020 nearly all secondary and undergraduate mathematics students were forced to adapt to new methods of learning and testing. As a result, the current experience for these mathematics students has changed vastly, as have the opinions and preferences of these students in terms of their learning. This study aims to identify instruction and testing resources and methods that students prefer and those that students find less beneficial in their current, post-pandemic educational experience. The past few years have seen a focus on the direct impacts of online learning during the pandemic, …

Collaborative Network Traffic Management Strategies Using The Distributed Reinforcement Learning And Large Language Models, Saeed Rashed Alkuwaiti

Thesis/ Dissertation Defenses

The focus of this research is to explore collaborative network traffic management strategies using the Distributed Reinforcement Learning (DRL) and Large Language Models (LLMs) approaches. It emphasizes exploring a new tool for addressing network traffic by utilizing Distributed Reinforcement Learning (DRL) and Large Language Models (LLMs). This is achieved by utilizing self-organizing and self-directing techniques to optimize the network performance. Using the NF-TON-IOT dataset, various classifiers such as Random Forest, AdaBoost, C4. 5, Multi-Layer Perceptron (MLP), and SVM with an RBF kernel were tested for traffic classification and intrusion detection. Research recommends that DRL optimizes the complexity of the network …

A Complete Guide To Multi-Fidelity No-U-Turn Sampling With Application To An Inverse Wave Problem, Paul Smith

Theses and Dissertations

We employ a Bayesian framework to solve an inverse problem that aims to identify the initial condition distribution of wave propagation, given a limited set of observable data. To optimize sampling efficiency, we introduce a multi-fidelity No-U-Turn Sampler (MF-NUTS) with an adaptive mass matrix. Maximizing sampling efficiency is crucial, as our ultimate goal is to extend this methodology to a large-scale inverse problem to identify the source parameters of historical tsunami-generating earthquakes. The forward model for tsunami propagation (GeoClaw) requires over 20 minutes of supercomputer time on 24 cores to evaluate each sample for a triple-fault rupture, necessitating a strategy …

Re: Comment Letter For The Butte Priority Soils Operable Unit (Bpsou) Butte Reclamation Evaluation System (Bres) Draft 2024 Corrective Action Plans (Dated January 27, 2025), Molly Roby

Silver Bow Creek/Butte Area Superfund Site

No abstract provided.

Gender Bias Within Ai Imaging, Drew Quattrocchi

Student Publications and Presentations

This study investigates AI-created gender bias in AI-created images through content analysis, contrasting the way gender is depicted in professions in leading AI image-creation tools such as Chat smith, Adobe Firefly, Midjourney, and Stable Diffusion. Employing a quantitative research method, this study contrasts AI-created images of gender-stereotypical careers for both male and female. Non-gendered careers will be used as well to identify patterns of stereotyping and bias. The area of emphasis lies in individual subjects within the images and scrutinizing visual elements such as clothing, accessories, background, face expressions, and gendered roles assigned to each. Particular emphasis is focused to …

Using Mathematical Modeling To Study The Dynamics Of Legionnaires’ Disease And Consider Management Options, Mark Z. Wang, Christina J. Edholm, Lihong Zhao

Faculty Articles

Legionnaires' disease (LD) is a largely understudied and underreported pneumonic environmentally transmitted disease caused by the bacteria \textit{Legionella}. It primarily occurs in places with poorly maintained artificial sources of water. There is currently a lack of mathematical models on the dynamics of LD. In this paper, we formulate a novel ordinary differential equation-based susceptible-exposed-infected-recovered (SEIR) model for LD. One issue with LD is the difficulty in its detection, as the majority of countries around the world lack the proper surveillance and diagnosis methods. Thus, there is not much publicly available data or literature on LD. We use parameter estimation for …