Physical Sciences and Mathematics Commons^™

Integrating Symbolic And Waveform Music Into Large Language Models, Teng Tu, Xiaohao Liu, Yunshan Ma, Ji Qi, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Music, as a unique and integral element of human life, is characterized by its complex structures, intricate details, and the fusion of multimodal information. Recent study advance music understanding by leveraging knowledge and reasoning capabilities derived from Large Language Models (LLMs). However, they often lack compatibility and fail to fully utilize the complementary strengths of diverse representations (e.g., ABC, MIDI, Waveform). To address these limitations, we propose a unified music-language model framework, named UniMuLM, transitioning from single-representation approaches to the integration of multiple music representations for LLM. Unifying different music representation formats poses challenges such as patch integrity and boundary …

Paid Search Marketing Vs. Search Engine Optimization: Analytical Models Of Search Marketing Based On Search Engine Quality, Kai Li, Chunyang Shen, Mei Lin, Zhangxi Lin

Research Collection School Of Computing and Information Systems

As search engines are leading revenue growth in online marketing, search marketing has become a popular area of academic research. Although search engine advertising has interested researchers for decades and much has been learned, one thing that puzzles scholars is why search engine optimization companies are tolerated rather than excluded from the market, even though they capture a significant share of the advertising market. In this paper, we shed light on this phenomenon and establish an analytical model based on organic search quality. Through analysis of the model, we were able to draw several intriguing conclusions. First, there is no …

Food Recognition With Visual Language Models: Search Re-Ranking Or Retrieval-Augmented Generation?, Kian Yu Gan, Phuong Anh Nguyen, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

Despite the rapid advances in Visual Language Models (VLMs), these models struggle to recognize culture-specific food items. While VLMs are effective in recognizing popular cultural dishes, their performance is suboptimal for dishes that are unique but not widely known internationally. Specifically, VLMs often generate either generic labels or hallucinated names for dishes that are localized to a particular culture. As a result, retrieval-augmented generation (RAG), which retrieves relevant recipes as references for VLMs, emerges as a promising approach. Nevertheless, recipe retrieval, which is itself imperfect, could mislead VLMs into generating inaccurate or culturally inappropriate dish names. This paper presents a …

Security-Enhanced Decentralized Conditional Privacy-Preserving Authentication In Vanets, Suqin Luo, Xinghua Li, Yinbin Miao, Xuelin Cao, Zhan Zhang, Yunwei Wang, Deng R.H.

Research Collection School Of Computing and Information Systems

To ensure the legitimacy of communicators while ad dressing the privacy concerns of vehicles in vehicular ad-hoc networks (VANETs), conditional privacy-preserving authentication (CPPA) schemes have been proposed. Given that existing schemes suffer from single point of failure due to centralized authorities, several distributed CPPA schemes have been proposed. However, these schemes all ignore the tight cementation between system secret keys and the authority, which could be a serious threat to system security, that the compromised authority may leak the system secret key. To address these issues, we propose a security enhanced decentralized conditional privacy-preserving authentication (DCPPA) scheme. DCPPA first introduces …

Clique Annealing: Semi-Supervised Community Detection Under Crystallization Kinetics, Ling Cheng, Jiashu Pu, Ruicheng Liang, Qian Shao, Hezhe Qiao, Feida Zhu

Research Collection School Of Computing and Information Systems

Semi-supervised community detection seeks to find a specified community type when only few communities are labeled. Existing "select-then-refine" pipelines often start from mis-aligned cores and rely on Reinforcement-Learning or Generative Adversarial Network, increasing computational cost and limiting scalability. We address these issues with a unified energy framework under crystallization kinetics that jointly models energy, structure, and growth. Based on this perspective, we propose CLique ANNealing (CLANN), which first employs Nucleus Proposer to select candidate clique as community core under four physics-inspired criteria. A learning-free Transitive Annealer then iteratively merges neighboring cliques and repositions the nucleus, enabling spontaneous, scalable community growth. …

Artem: Enhancing Large Language Model Agents With Spatial-Temporal Episodic Memory, Cassandra Hui Ming Tan, Budhitama Subagdja, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

Current large language models (LLMs) exhibit significant deficiencies in episodic memory tasks including encoding, storing, and retrieving specific information from temporally dependent events over a long period of time. Recent approaches to handle memory tasks in LLMs, such as in-context learning, retrieval-augmented generation (RAG), and fine-tuning, may resolve the long-term retention issues, but are still inadequate to handle tasks requiring chronological awareness of the stored information. We introduce Agentic Retrieval with Temporal-Episodic Memory (ARTEM), a hybrid LLM-based agent architecture integrating LLMs with a self-organizing neural network named Spatial-Temporal Episodic Memory (STEM), designed to handle episodic memory tasks. Our approach employs …

Dystop: Dynamic Staleness Control And Topology Construction For Asynchronous Decentralized Federated Learning, Yizhou Shi, Qianpiao Ma, Yan Xu, Junlong Zhou, Ming Hu, Yunming Liao

Research Collection School Of Computing and Information Systems

Federated Learning (FL) has emerged as a potential distributed learning paradigm that enables model training on edge devices (i.e., workers) while preserving data privacy. However, its reliance on a centralized server leads to limited scalability. Decentralized federated learning (DFL) eliminates the dependency on a centralized server by enabling peer-to-peer model exchange. Existing DFL mechanisms mainly employ synchronous communication, which may result in training inefficiencies under heterogeneous and dynamic edge environments. Although a few recent asynchronous DFL (ADFL) mechanisms have been proposed to address these issues, they typically yield stale model aggregation and frequent model transmission, leading to degraded training performance …

Realign: Text-To-Motion Generation Via Step-Aware Reward-Guided Alignment, Wanjiang Weng, Xiaofeng Tan, Junbo Wang, Guo-Sen Xie, Pan Zhou, Hongsong Wang

Research Collection School Of Computing and Information Systems

Text-to-motion generation, which synthesizes 3D human motions from text inputs, holds immense potential for applications in gaming, film, and robotics. Recently, diffusion-based methods have been shown to generate more diversity and realistic motion. However, there exists a misalignment between text and motion distributions in diffusion models, which leads to semantically inconsistent or low-quality motions. To address this limitation, we propose Reward-guided sampling Alignment (ReAlign), comprising a step-aware reward model to assess alignment quality during the denoising sampling and a reward-guided strategy that directs the diffusion process toward an optimally aligned distribution. This reward model integrates step-aware tokens and combines a …

Airaclex: Automated Detection Of Price Oracle Manipulations Via Llm-Driven Knowledge Mining And Prompt Generation, Bo Gao, Yuan Wang, Qingsong Wei, Yong Liu, Rick Siow Mong Goh, David Lo

Research Collection School Of Computing and Information Systems

Decentralized finance (DeFi) applications depend on accurate price oracles to ensure secure and fair transactions. However, poorly integrated oracles remain susceptible to manipulation, enabling attackers to exploit smart contract logic for unfair asset valuation and financial gain. While many such vulnerabilities are only detected after deployment, smart contracts are typically immutable once deployed, making post-hoc fixes costly or infeasible. This highlights the critical need for detecting oracle manipulation risks before deployment. In this paper, we propose AiRacleX, a novel LLM-driven framework that enables pre-deployment detection of price oracle manipulation vulnerabilities by leveraging the complementary strengths of multiple large language models …

Bayesian Modelling On Periodically And Multiple Periodically Correlated Time Series Data, Jie Yao

Electronic Theses & Dissertations (2024 - present)

Time series with multiple periodically correlated (MPC) components present a complex challenge, with relatively limited prior research. Most existing models are designed for simpler periodically correlated (PC) components and often struggle with over-parameterization, optimization issues, and capturing complex PC patterns within a time series. Frequency separation techniques can help preserve the correlation structure of individual PC components, while Bayesian methods can integrate new and prior information to refine beliefs about these components. This study proposes a two-stage approach that combines frequency separation and Bayesian techniques to forecast PC and MPC time series data. This method aims to demonstrate improved effectiveness …

Bio-Orthogonal Chemistry-Based Strategy To Turn-On And Turn-Off Crispr-Cas9 Gene Editing In Solution And In Live Cells, Bhoomika Pandit

Electronic Theses & Dissertations (2024 - present)

The CRISPR–Cas9 system is a widely popular tool for genome engineering. There is a strong interest in developing tools for temporal control of CRISPR-Cas9 activity to address some of the challenges and to broaden the scope of potential applications. In this thesis I describe two biorthogonal based approaches to either Turn-ON and Turn-OFF CRISPR-Cas9 gene editing.

In first project work I describe a bio-orthogonal chemistry-based approach to Turn-ON Cas9 nuclease activity with temporal precision. We report a TCO-acylimidazole reagent that acylates 2′-OH groups of RNA. Poly-acylation (“cloaking”) of RNA was optimized in vitro using a model 18-nt oligonucleotide, as well …

Applications Of Machine Learning For Evaluating Downward-Coupled Stratosphere-Troposphere Interactions And Subseasonal Forecasts Of Opportunity, Elena M. Fernandez

Electronic Theses & Dissertations (2024 - present)

Wintertime stratospheric dynamics provide key information for understanding atmospheric teleconnections and improving subseasonal-to-seasonal (S2S) predictions on timescales of two weeks to two months. Periods of enhanced predictability, often referred to as forecasts of opportunity, arise from large-scale teleconnected variability, within which the stratosphere serves as an important precursor for tropospheric states, such as near-surface temperatures. While traditional diagnostics of downward coupled stratosphere-troposphere interactions typically rely on zonal-mean representations of wind and geopotential height, this dissertation presents an alternative vortex-centric framework through metrics that capture the daily geometric and dynamical evolution of the stratospheric polar vortex. The proposed stratospheric …

Investigation Of Fine-Grain Cu And Cu Alloys For Low-Temperature Hybrid Bonding Applications, Sarabjot Singh

Electronic Theses & Dissertations (2024 - present)

Hybrid bonding has emerged as a key enabler for next-generation three-dimensional (3D) integration, offering fine-pitch interconnects and improved electrical performance. However, conventional Cu–Cu hybrid bonding typically requires elevated temperatures to achieve sufficient diffusion and interface quality, posing challenges for temperature-sensitive device integration and process compatibility. This work investigates materials engineering approaches to enable low-temperature Cu–Cu bonding through both microstructure design and alloying strategies.

This work begins by examining grain refinement in Cu as a pathway to enhance diffusion through increased grain boundary density, providing efficient atomic transport without introducing additional elements. Three Cu-based systems Cu–Co, Cu–Ag, and Cu–Al were systematically …

Photo And Spatial Control Of Rna Structures And Functions: Exploring Positional Effects And Photo-Responsiveness Of Azobenzene Modified Nucleotides, Jinxi Du

Electronic Theses & Dissertations (2024 - present)

RNA molecules perform many important biological functions by forming complex three-dimensional structures stabilized by base pairing, tertiary interactions, and metal ion coordination. Because RNA function depends strongly on its structure, methods that allow reversible control of RNA conformation and activity are important for understanding RNA behavior and for potential therapeutic applications. However, achieving reversible and site-specific control of RNA structure, especially at the single-nucleotide level, remains challenging. In this dissertation, I developed and applied azobenzene-modified nucleotides as a chemical tool to enable optical and spatial control of RNA structure and function.

The first part of this dissertation focuses on establishing …

Hidden Markov Models For Stock Market Regime Prediction, Meghan Blanch

Electronic Theses & Dissertations (2024 - present)

Hidden Markov Models (HMMs) are a powerful mathematical system used for analyzing time sequences in which the underlying states are unable to be directly observed. This project strives to develop the theory and computation of such models, including deriving the forward and backward variables, the Baum-Welch algorithm, and the Viterbi algorithm. Each of these HMM components are derived and explained to demonstrate the probabilistic principles that allow HMMs to be effective for modeling sequential data.

To demonstrate its practical relevance, the HMM is applied to financial time series. The observable stock market returns tend to be influenced by market regimes …

Scalable Single-Erbium Telecom Qudits With Record Room-Temperature Quantum Coherence In Silicon-Based Nanostructures, Alexander Kaloyeros

Electronic Theses & Dissertations (2024 - present)

Advancing quantum information science demands solid-state quantum systems that maintain long quantum coherence at elevated temperatures while supporting scalable, CMOS-compatible fabrication and telecom C-band operation. No existing platform has simultaneously achieved these requirements, as state-of-the-art demonstrations of coherent control of erbium ions, with an intrinsic telecom-band optical transition, have been confined to cryogenic temperatures below < 10 K under controlled vacuum conditions. This thesis introduces a new paradigm in which materials science and engineering provides the enabling pathway to quantum coherence.

A foundry-compatible nanofabrication approach, paired with targeted materials engineering, is developed to realize a new class of CMOS-scalable quantum system: arrays of spatially isolated single-erbium-ion qudits (five-level systems) embedded in silicon-based (e.g., silicon carbide (SiC) and SiC_xO_y) hollow nanopillars (HNPs). Non-lithographically …

Historical Changes In The Seasonality And Variability Of Daily Surface Air Temperature In Era5 And Cmip6 Models, Zirui Wan

Electronic Theses & Dissertations (2024 - present)

This thesis investigates historical changes in the seasonality and daily variability of near-surface air temperature (Tas) during 1950–2024 using ERA5 reanalysis and simulations from 21 CMIP6 models. ERA5 and CMIP6 show broadly similar climatological structures of the Tas seasonal cycle and variability, with considerable inter-model discrepancies. Both ERA5 and CMIP6 show decreasing Tas variability over northern mid-high latitudes, while Tas variability increases in the low latitudes in ERA5 but changes little in CMIP6 multi-model mean (MMM) due to different change patterns among the models. To investigate the mechanisms of historical Tas variability changes, I analyzed the relationships between Tas variability …

The Chemistry Of Red Lipstick: The Impact Red Lipstick Has On Society, Lilliana Weldeslassie

Electronic Theses & Dissertations (2024 - present)

Cosmetic chemistry is deeply embedded in everyday life, shaping the products people use from morning to night. Among these, red lipstick stands out as a formulation that blends complex chemistry with a cultural meaning. This thesis examines the scientific foundations of red lipstick through its chemical composition, pigment structure, toxicological history, analytical methods, and evolving sustainability practices. Historically, red lipsticks relied on high-risk ingredients such as lead, mercury, and other toxic metals, reflecting an era before chemical safety and regulatory oversight were incorporated.²¹ Modern formulations have established safer synthetic dyes, natural waxes, plant-based oils, and stabilizers created to optimize …

Informationally-Optimal Measurements On Single-Qubit Systems, Adam V. Preston

Electronic Theses & Dissertations (2024 - present)

This thesis will focus on deriving informationally-optimal quantum state tomographic measurements on single-qubit systems, with the definition of informationally optimal to be defined as those measurements which maximize the average information gain. The informationally-optimal measurements that will be covered include projective measurements (formally building off of the work of [1] and putting it on firm, information-theoretic foundations), and more general types of quantum measurements called directional (also known as rank-one) positive operator-valued measure (POVM) measurements, and, finally, adaptive directional POVM measurements.

For projective measurements, we build on the work of Wootters and Fields ([1]) and show, via analytical methods and …

The Effects Of Downdraft And Radial Ventilation In Hurricanes Earl (2010) And Edouard (2014), Jacob Vile

Electronic Theses & Dissertations (2024 - present)

Tropical cyclone (TC) intensity change in moderate vertical wind shear (VWS) is difficult to predict. Specifically, ventilation, the intrusion of dry and/or cool (i.e., low equivalent potential temperature (theta-e) air inside the TC, can have differing effects on TCs. Downdraft ventilation is the downward transport of below-mean  theta-e air by downdrafts, and radial ventilation is the inward transport of below-mean theta-e air by storm-relative radial winds. If this low theta-e air is able to get ingested by updrafts, it can result in a reduction of convection in the TC, thereby reducing intensity. Although most studies of ventilation have been conducted …

Key Processes Involved In The Vortex Alignment Of Sheared Tropical Cyclones, Luis O. Hernandez

Electronic Theses & Dissertations (2024 - present)

Tropical cyclones (TCs) are challenging to forecast when under vertical wind shear (VWS). VWS tilts a TC’s vortex, creating asymmetries in its structure that impedes intensification. Although work has been done to conceptualize and understand how TCs evolve under shear and how they axisymmetrize and intensify, vortex alignment remains a complex and unresolved topic. The goal of this study is to assess what physical processes are important to tilt reduction and how these processes evolve in weak TCs. Using the Hurricane Analysis Forecast System Model-B (HAFS-B) 2.1, a case study on two storms, Hurricane Earl (2022) and Hurricane Lee (2023) …

Drivers Of Extreme Streamflow During A Snowmelt-Enhanced Heavy Rainfall And Atmospheric River Event In The Catskill Mountains Of New York, Christopher A. Gilberti

Electronic Theses & Dissertations (2024 - present)

This study uses atmospheric reanalysis, precipitation and snow water equivalent (SWE) analyses, mesonet observations, and streamflow records to examine the impact of an atmospheric river storm that produced heavy rainfall and snowpack ablation across the Catskill Mountains of New York on 24–25 December 2020. Earlier in December, an antecedent storm produced SWE values in the Catskills above the 90^th percentile for that point in the season. The atmospheric river event followed with strong southerly warm-air advection ahead of an anomalously deep trough, temperatures and dewpoints that exceeded 10°C, and heavy orographically enhanced rainfall and rapid snowmelt. Sensible heat fluxes …

The Impact Of Downdraft Ventilation On Tropical Cyclone Intensity And Structure Using Aerial Reconnaissance Observations, Nicholas E. Johnson

Electronic Theses & Dissertations (2024 - present)

Tropical cyclone (TC) intensity can be challenging to predict since it is modulated by many physical processes. One of these processes is ventilation: the injection of relatively cool, dry environmental air into the warm core of a TC that acts as “anti-fuel”, inhibiting intensification and increasing uncertainty with intensity forecasts. While ventilation has been examined in numerical modeling studies, there have been a limited number of observational studies investigating ventilation. The two primary objectives of this dissertation are to develop methods for measuring downdraft ventilation from aerial reconnaissance observations and to use those observations to better understand how ventilation affects …

Utilizing Coal For Mesophase Pitch-Based Carbon Fiber Production: Precursors, Processes, And Progress, Christina M. Thompson

Theses and Dissertations--Chemistry

Graphitic materials possess unique properties due to the unique combination of layered crystalline structure and carbon’s low atomic weight. High performance carbon fiber is one such example, displaying exceptional strength-to-weight ratios, stiffness, and thermal and chemical resistance. These properties render high performance carbon fiber a critical structural reinforcement material in the manufacture of composites across various industries, such as for automotive and aerospace applications. However, balancing fiber performance with precursor and processing costs remains a challenge. As alternative carbonaceous feedstocks are explored, coal has gained interest for utilization in graphitic products as a relatively abundant and low-cost source of aromatic …

Elucidating The Impacts Of Non-Covalent Interactions In Organic Materials Through A Multiscale Computational Approach, Sashen A. Ruhunage

Theses and Dissertations--Chemistry

Noncovalent interactions (NCIs) in π-conjugated organic materials serve as tunable levers that influence molecular structure and intermolecular interactions in the condensed phase and, in turn, impact the electronic, optical, and mechanical properties of these materials. NCIs include attractive dispersion, electrostatic, and induction interactions, as well as repulsive exchange interactions. However, how to design materials with NCI considerations remains an open question across many fields. Here, we seek to provide an electronic and atomistic perspective on these interactions through multiscale simulations to aid materials design, processing, and performance optimization. In this study, we investigate NCIs and their effects across various systems …

Student Familiarity With The Periodic Table Of The Elements: Results From Cued-Recall And Eye-Tracking Assessments On Memory, Victor A. Okuo

Theses and Dissertations--Chemistry

Learning element symbol–name associations and the spatial organization of elements on the periodic table is a foundational step in learning chemistry, supporting later understanding of chemical formulas, equations, bonding, and stoichiometry. Although students often rely on memorization strategies to learn periodic table content, this task is challenging due to the large number of elements and the apparent ambiguity in matching some element symbols to their names.

This study explores students’ recall of element names when given element symbols as cues and their knowledge of element locations on the periodic table when given element names as cues. The study also examines …

Symbolic Logistic Regression For Interval-Valued Predictors: A Simulation Study And Application To Health Data, Soad Abdullah

Graduate Research Theses & Dissertations

This thesis investigates symbolic logistic regression for interval-valued predictors through simulation studies and a real health data application. Classical logistic regression assumes exact predictor values, whereas in many practical settings variables are available only in interval form due to coarsening or reporting uncertainty. Two simulation studies examine the impact of interval uncertainty under asymmetric intervals and measurement error. Symbolic models based on midpoint and midpoint-plus-width representations are compared with the classical approach. Results show that midpoint modeling captures the general relationship but introduces bias under asymmetry, while incorporating width reduces this distortion. Under measurement error, classical logistic regression exhibits attenuation …

Plant Practices And Community At The Heart Of Hopewell: Paleoethnobotany At The Overly Site And Beyond, Abigail Deewaard

Graduate Research Theses & Dissertations

Paleoethnobotanical analyses of Ohio Hopewell habitation sites in past decades have shed light on the subsistence practices, settlement patterns, and environmental management of Indigenous peoples engaged in what archaeologists describe as the Hopewell phenomenon. However, comparatively less attention has been paid to the social dimensions of plant use, particularly at habitation sites in the Scioto Hopewell region. Hopewell is an archaeological term that refers to the Middle Woodland (ca. 100 BCE–400 CE) network of interregional relationships across eastern North America characterized by long-distance exchange, monumental earthwork construction, and ceremonial practices including the honoring of ancestors. This thesis applies a social …

New Methods For Bright Electron Beam Generation At The Argonne Wakefield Accelerator Facility, Emily Frame

Graduate Research Theses & Dissertations

Bright electron beams generated via photoinjectors have enabled many recent developments in accelerator technology, including x-ray free electron lasers and ultrafast electron microscopy and diffraction systems. The conceptual design, scientific reach, and accessibility of these and next-generation accelerator applications can be significantly improved by increasing electron beam brightness. Therefore, the development of novel methods to further enhance beam brightness remains a critical priority in the field.

One of the primary factors limiting beam brightness is the mean transverse energy (MTE) of photoemitted electrons. The MTE is an intrinsic property of the photocathode and is proportional to the square of the …

4d-Qens And Diffuse Scattering: Analysis Of Dynamics And Local Ordering In Superionic Conductors, Jared Coles

Graduate Research Theses & Dissertations

Ionic conductors are increasingly central to electrochemical energy technologies because they enable charge transfer through the motion of mobile ions. Among them, superionic conductors are crystalline solids that exhibit exceptionally high ionic conductivities, making them promising candidates for safer and more sustainable next‑generation energy storage and conversion devices. Despite their technological importance, key gaps remain in our ability to connect macroscopic transport to microscopic structural disorder and correlated dynamics that enable fast‑ion motion. Neutron scattering provides a powerful set of tools for establishing these connections.

In total scattering experiments, Bragg diffraction measures the average crystallographic structure, while diffuse scattering probes …