Scholar Commons - Institutional Repository of the University of South Carolina
Not a member yet
34867 research outputs found
Sort by
Knowledge and Ontology Enhanced Approach to Natural Language Understanding (KOE-NLU) in Computational Social Media and Healthcare
No full textNatural Language Understanding (NLU) faces both opportunities and challenges as the amount of social media and healthcare data grows. This is particularly evident in context-sensitive applications such as evaluating cognitive health, identifying mental health symptoms, and monitoring drug abuse. Even though traditional NLU models work well for processing language in a wide range of areas, they often lack the ability to understand language in a specific domain, reason in context, and incorporate structured external knowledge. This dissertation talks about the Knowledge and Ontology Enhanced Approach to Natural Language Understanding (KOE-NLU), a new framework that is meant to make NLU systems better at understanding semantic depth, contextual awareness, and drawing conclusions in computational social media and healthcare informatics. The KOE-NLU framework is built upon three fundamental pillars: There are three types of methods. The first is Ontology-Driven Semantic Representation, which puts domain-specific knowledge into formalized ontologies to help with making sense of things based on their context and drawing conclusions from them. The second is Knowledge Graph Integration, which links different types of data sources to improve structured representation and reasoning over unstructured text. The third type is transformer-based language models with knowledge augmentation. This type of modeling uses domain-specific prompts to teach transformer based models, generative AI models structured knowledge sources. These components work in synergy to enable more accurate and context-aware natural language predictions. To demonstrate the efficacy of the KOE-NLU framework, this dissertation presents four major application areas. In the first place, the framework is used in cognitive health informatics to predict Mild Cognitive Impairment (MCI) by putting together MoCA test scores and automated discourse analysis. This is done by using a Cross-Cognitive Domain Attention (CCDA) model to pull out linguistic markers that show cognitive decline. Second, we present process knowledge-guided language prompting as a way to automate Main Concept Analysis (MCA) for discourse assessment. It is shown that process-aware language models are better than traditional text-based classifiers at detecting speech coherence in people with neurodegenerative conditions. Third, in mental health analytics, a knowledge-infused multi-task learning framework is developed to extract mental health symptoms linked to cardiovascular disease (CVD) from social media discourse, employing hierarchical attention networks combined with expert-curated knowledge bases. Finally, the dissertation introduces a Drug Abuse Ontology (DAO) that was created using semi-automated ontology engineering methods. We can use this ontology to identify p atterns in substance use disorders, monitor illicit drug trends on social media, and examine the evolution of drug-related emotions over time. A rigorous experimental framework is implemented to evaluate the KOE-NLU models across these applications. The role of ontology-enhanced reasoning in model performance is looked at by comparing supervised and self-supervised learning paradigms. The results show that knowledge-infused transformer architectures do better than baseline deep learning models and are better at interpreting clinical discourse in terms of context. Structured ontological constraints also make it much easier to classify substance use disorders and pull-out mental health symptoms. This dissertation makes computational social media analytics and healthcare informatics better by giving us a knowledge-driven NLU framework that can be used on a large scale. This framework connects data-driven machine learning with symbolic AI methods. The KOE-NLU framework creates the foundation for the next generation of AI models that can be explained. These models will use process knowledge, ontological reasoning, and knowledge graphs to help computers understand language better in important healthcare settings. More work can be built on top of this by including speech, physiological, and neuroimaging signals, making domain-adaptive self-learning ontologies, and setting up real-time clinical decision support systems. This research marks a significant advancement in KOE-NLU, with important real-world applications in healthcare AI, clinical decision support, and public health monitoring on social media
House Rules: Athlete, Coach, and Staff Perspectives on the Shifting NCAA Landscape
No full textThe House settlement is an inflection point in NCAA Division I athletics, ushering in a new model of institutional revenue sharing, roster regulation, and athlete compensation. While legal and financial analyses dominate public discourse, less is known about how these shifts are understood by those working and competing within athletic departments. This study draws on narrative analysis of interviews with 11 athletes, coaches, and staff across five anonymized Power 4 institutions to examine how the House settlement is being interpreted as it takes effect. Three narrative foci emerged: 1) financial prioritization and resource allocation, 2) shifting culture, relationships, and meaning, and 3) organizational and bureaucratic (non)response. Participants described uneven access to information, evolving understandings of roles and responsibilities, and deepening stratification across programs. While some viewed the settlement as formalizing existing trends, others expressed disorientation, concern, or loss. This study offers one of the first qualitative accounts of the settlement’s lived impact and perceived institutional implications by centering these early interpretations
Transgressive Desire and Women’s Liberation in the British Strike \u3ci\u3eBildungsroman\u3c/i\u3e
No full textThis dissertation examines the intersection of gender and class identity in British proletarian Bildungsromane set during the General Strike of 1926 and the Miners’ Strike of 1984/5, referred to collectively as “Strike Bildungsromane.” These prolonged industrial conflicts temporarily disrupted the gender division of labor that had hitherto been a defining aspect of pit village life. This development encouraged proletarian writers of the interwar and post-war periods to consider coalfield women as political agents, capable of subverting their preordained roles within the private sphere. These moments of subtle subversion are critical to the General Strike novels explored in this study, including Lewis Jones’s We Live (1939), Ellen Wilkinson’s Clash (1929), and Harold Heslop’s The Gate of a Strange Field (1929). These tensions resurface more overtly throughout such Miners’ Strike Bildungsromane as Kit Habianic’s Until Our Blood Is Dry (2014) and Janet MacLeod Trotter’s Never Stand Alone (1997). Using both symptomatic reading (Jameson, Macherey) and Marxist-feminist theory (Federici, Hartsock, Hartmann), the study interrogates the affordances and limitations of the class-based ontology underlying the proletarian Bildungsroman. While General Strike Bildungsromane in particular manage to grant visibility to coalfield women’s roles within the home, these duties are still largely treated as leading to an absence of political consciousness. The study concludes that deindustrialization during the Thatcher era forced proletarian writers to reevaluate the primacy of class identity, thereby granting pit village women a political identity essential to, but apart from, the masculinist class struggle
The Return Journey: Migration, Reintegration, and Postcolonial Identity in Contemporary African Literature
No full textMigration is a human practice as old as human society. It is primarily triggered by a human urge for self-preservation or the quest for better standards of living. Various periods of human history have known great flows of voluntary, semi-voluntary, and forced migrations. In contemporary days, people move more than ever before, mostly from countries of the Global South to countries of the Global North, mainly from African countries. These migrants flee from poverty, armed conflicts, unemployment, and repression from despotic leaders. While the question of migration has consistently and extensively been addressed, the focus in the mainstream migration studies has always been on the movement from the home country to a host country, neglecting return migration, which, despite being less intensive, remains important for fully understanding the human migration experience. Focusing on Cheikh Hamidou Kane’s novel Ambiguous Adventure and Imbolo Mbue’s How Beautiful We Were, this thesis analyzes the migration and return journey of two protagonists, drawing from Homi Bhabha’s postcolonial notions of hybridity and estrangement. The main argument of this thesis is that the return journey is similar to the migration to a new country. Both require the subject to undergo a re/integration process to be able to be socially active
The Other Side of Trusting: The Implications of Vulnerability as a Core Component of Organizational Relationships
No full textTrusting behavior—discretionary risk-taking with another party—is the primary means of facilitating productive workplace relationships (Zand, 1979; Mayer, Davis, & Schoorman, 1995). Yet, little empirical evidence exists to explain what happens after employees engage in trusting behavior, but before outcomes materialize—otherwise termed the “vulnerability phase” (Ballinger, Schoorman, & Sharma, 2024: 2). To address this neglect, I develop a theoretical framework that positions vulnerability as a core psychological state within the trust process, shaped by trustors’ cognitive appraisals of risk severity and uncertainty. Drawing from research on stress and well-being, the framework also identifies coping as a key response to vulnerability, moderated by various contextual features. Utilizing a four-wave survey of 636 Army servicemembers, I test a moderated-mediation model proposing that an employee’s experience of vulnerability, generated from their trusting behavior, leads to unintended, negative consequences. The results find modest support for the model
Saw Biosensor for the Detection of Cyanobacteria and Cyanotoxin
No full textThe growing prevalence of cyanobacterial blooms and associated cyanotoxins, such as microcystin-LR (MC-LR), presents a significant threat to water quality and public health. Conventional detection methods, including ELISA and PCR, are often time-consuming, expensive, and require trained personnel, limiting their application in field settings. Hence, a miniaturized, quick testing sensor is proposed to circumvent the current sensing challenges for real time, in-situ diagnostics of MC-LR. The proposed high-frequency ultrasonic sensor employs surface acoustic waves (SAW). The sensor was designed using a multi-physics simulation tool to understand wave propagation in the piezoelectric substrate (lithium tantalate) and optimize the design of the electrodes to achieve resonance. Once the design was finalized, the sensors were fabricated to trigger and maintain shear-horizontal Love waves on lithium tantalate wafers. After fabrication, the sensors were functionalized with MC-LR antibodies following a specific, previously established protocol. To evaluate the functionality of the sensors, water samples were collected from stormwater and agricultural ponds in South Carolina. Each of these specimens were then assessed across six sensors with the designed electrodes. For each sample and each sensor, time-domain ultrasonic wave signals were collected at a resonance frequency of 12 MHz at 10-minute intervals after adding the samples. Signals were analyzed using a multi-fidelity feature extraction tool. Results demonstrated an MC-LR concentration-dependent frequency shift with a conservative limit of detection estimated at 5 μg/L. The sensor’s selectivity for MC-LR antigens over unrelated antigens was previously shown. Reproducibility was observed across all biosensors, though variability increased at higher concentrations, possibly due to environmental interference or progressive cyanobacteria lysis. This SAW biosensor shows promise as a rapid and portable tool for MC-LR detection. Future work will focus on sensor miniaturization, packaging refinement, and improved calibration across mid-range concentrations
Episode 88: Sixty Years of Transformation: Harry Lesesne on the Making of a Modern University
No full textMore than 25 years ago, Harry Lesesne was a doctoral history student at USC, tasked with writing a new history about the university that would cover the years 1940 to 2000. He recounts how he took on the assignment to narrate the six decades that transformed Carolina into a modern research university
A/r/tography: Looking Back to Look Forward
No full textThis presentation will provide a review of four themes that have dominated a/r/tographic literature while providing an example for each: relationality and renderings; ethics and embodiment; movement and materiality; and propositions and potentials. In so doing, this talk provides an introduction to individuals interested in a/r/tography as a potential practice while assisting those who are already using a/r/tography to contemplate new pursuits
C \u3csup\u3e3\u3c/sup\u3e AN: Custom, Compact and Composite AI Systems - A NeuroSymbolic Approach: 4\u3csup\u3eth\u3c/sup\u3e-Generation Evolution of Intelligent Systems
Get PDFArtificial Intelligence (AI) systems continue to evolve rapidly. From the architecture perspective, it is evolving from large, monolithic models trained on massive internet data to complex, multi-component “compound” systems and “agentic” frameworks capable of semi-autonomous decision-making. These systems show immense promise yet face numerous challenges in reliability, consistency, transparency, and alignment with user goals. In this article, we propose Custom, Compact and Composite AI with Neurosymbolic (C3AN) approach, a framework that paves way to 4th-generation of AI that integrates data, knowledge, and human expertise to build robust, intelligent and trustworthy AI systems defined by 14 foundation elements.
Custom emphasizes the focus on high-quality, domain-specific data and knowledge, along with tailored workflows and user or application-specific constraints. Compact highlights resource-conscious implementation that does not require extreme scale to achieve reliable domain adaptation. Composite refers to the integration of multiple AI modules that collaboratively perform domain-specific tasks, handling data, knowledge, and human expert feedback within a cohesive Neurosymbolic framework. Together, these qualities address longstanding issues in large, monolithic, or purely black-box models. We illustrate the foundation elements of C3AN in two complex AI systems with demands representative of enterprise class and/or mission critical applications: (1) Nourich, a disease-specific diet management system that recommends recipes based on users’ health condition and food preferences, and (2) MAIC (MTSS AI Concierge), which operates in the Multi-Tiered System of Supports (MTSS) domain for mental health and behavioral interventions to support health workers with different roles. We conclude by outlining practical challenges and future research directions to foster robust, multi-domain adoption of C3AN
Attendance and Absenteeism: Can We Fix It?
Get PDFThis research project shows how consistent communication with parents of special education students regarding daily and weekly attendance can make a difference in decreasing absenteeism and increasing attendance in the virtual resource classroom. This study supports promoting access to special education services with parents as partners in the virtual environment. This approach can be adapted to fit various classroom environments