Doctoral Dissertations

Permanent URI for this collection

https://hdl.handle.net/10919/11041

Browse

Now showing 1 - 20 of 18083

Explainable AI for Social Good: Applications in Mental Health, Public Health Risk, and Environmental Traceability
Sarkar, Shailik (Virginia Tech, 2026-01-12)
The ubiquitous use of machine learning and AI technology in human-centered domains such as social networks, public health, sustainable trade, and environmental forensics indicates a significant need for an adaptive, interpretable, and generalizable approach in predictive modeling. With the increasing availability of user-generated data, environmental samples, and public health records, AI-driven tools have played a significant role in predictive analysis. However, a persistent challenge remains: in domains with significant societal implications, the availability of data is often inconsistent, unstructured, and lacks fine-grained labels. Furthermore, in these application areas, understanding the prediction becomes as important as the prediction itself, as they guide a more informed intervention strategy. Most of the existing work in this domain struggles to meet this requirement by approaching it from either one size fits all modeling approach or by adapting to a very problem-specific, fine-tuned algorithm that fails to learn the inter-task dependency while not particularly focusing on explainability. Therefore, in real-world scenarios, these tools show an increased risk for practical applicability due to their black-box nature, which leads to a lack of intuitive interpretability for domain experts. These methods can often neglect underlying conditions such as spatial dynamics, socioeconomic disparities, and uncertainty. In fields like population health management and stable isotope forensics, such limitations hinder practical deployment and erode trust. Compounding this issue is the widespread adoption of large language models (LLMs), which, despite their power, are prone to hallucinations and toxicity, undermining their reliability in sensitive domains.This thesis employs active learning, multi-task learning, ante-hoc explainability, post-hoc explanations, and probabilistic Gaussian process modeling to tackle several domains of social computing that range from population mental health, epidemiological outbreak, and forensic environmental tracability analysis. The first work introduces AMMNet, a multi-task active learning model for detecting depression and anxiety from Reddit data. It combines topic-based embeddings and joint task training to improve interpretability and data efficiency over conventional LLM-based classifiers. The main contributions are: 1. It tackles the lack of a fine-grained labeled dataset for Reddit that extends beyond topic-specific subreddits by first curating a labeled dataset and then employing an active learning strategy to help with the training; 2. It proposes a novel multi-task learning model, AMMNet, that outperforms baseline models in the prediction of mental health conditions. 3. A novel model-level explanation behind our prediction due to the introduction of the task-specific feature selector in the task-specific module; and 4. It shows through extensive experiments that for domain-specific classification tasks such as this, a combination of document-level embedding and topic distribution gives the best performance across all the tasks. In the second work, DeMHeM, a multi-task model for identifying bipolar disorder and its comorbidities, is introduced. Through soft parameter sharing and focal loss, the model robustly detects nuanced mental health states and facilitates deeper community-level insight via keyphrase analysis. The main contributions are: 1. development of a novel multitask learning framework for mental health predictions; 2. implementation of a novel and effective multitask optimization algorithm; and 3. exploring post-hoc analysis using the trained model for a more fine-grained understanding of bipolar disorder and its comorbidity. The third work proposes GC-Explainer, an explainable Graph Neural Network for forecasting COVID-19 outbreak severity using only static population features. The model integrates explainability directly into its architecture, enabling transparency without post-hoc methods, and avoids the reliance on real-time or temporal data. The main contributions of the work are: 1. Unlike post-hoc methods for GNN explanation, this work proposes a novel framework, Graph-Covid-Explainer, that simultaneously gives predictions for high-risk areas as well as insights about the most important features during the training of the model. 2. It introduces a novel problem setting that tackles the paucity of historical data to identify high-risk areas during the initial outbreak that can help authorities in better preparing for future crises, and 3. it applies Graph-COVID-Explainer(GCExplainer) on real-world COVID-19 data to show that static features about mobility, socioeconomic status and spatial dependency among regions can be used to make an explainable prediction about the varied degree of severity during the early part of the outbreak, without using historical pandemic data as features.The fourth work proposes to deliver a deployed pipeline that combines Stable Isotope Ratio Analysis (SIRA) and environmental variables using a multi-task Gaussian Process framework. It provides origin tracing for timber samples with predictive uncertainty, significantly improving upon traditional spatial regression approaches. The main contributions are: (1) It presents a comprehensive multi-task Gaussian process modeling framework that supports the incorporation of auxiliary data, such as climate layers, to support origin determination. This enables the incorporation of environmental factors, imputing uncertainty to predictions, and multimodal feature integration; (2) This work is a deployed machine learning pipeline wherein physical samples are collected, subject to tests, and injected into our model to help European enforcement agencies in combating illegal timber trade by demonstrating that a claimed harvest location other than Russia is not viable; and (3) It demonstrates accuracy profiles of our approach in a controlled experiment that illustrates the interplay between SIRA values and atmospheric variables and how they affect our ability to reveal harvest location misrepresentation. This goes beyond traditional ML pipelines that only predict isotope values into an end-to-end approach that supports decision-making by enforcement agencies. The final work combines the concepts of matrix sparsification to extract feature importance with epistemic uncertainty arising from Gaussian processes to make explainable spatial prediction of health outcome in the form of type-2 diabetes. Through a rigorous experimental design, the novel end-to-end Machine Learning framework Deep Graph Gaussian Health Net(DDHG-Net) demonstrates the effectiveness of the model compared to state-of-the-art across different metrics while also providing feature-level insights and uncertainty-aware prediction, making it more suitable for real-world applicability. Our case study on Virginia demonstrates this effectively by identifying a highly prevalent cluster of counties more accurately with high confidence, while uncertain predictions also give insight about which geographical area should conduct a more careful data collection. Overall, the proposed methodological approach laid out in this dissertation promises to be effective in different real-world application domains where explainability is paramount, and the immediate impact of these works lies in greater community welfare.
Reclaiming the Appalachian Landscape: Understanding Surface Mine Reclamation Through Remote Sensing
Putnam, Daniel Jacob (Virginia Tech, 2026-01-12)
Surface mining has been the dominant driver of land cover change in the Central Appalachian region for decades, fundamentally altering the topography and ecological trajectory of over one million acres of forest. While the Surface Mining Control and Reclamation Act (SMCRA) of 1977 mandated the restoration of mined land to an equal or better land use, the ecological integrity of these reclaimed landscapes remains uncertain, often characterized by "arrested succession" and the proliferation of invasive species. This dissertation integrates multi source remote sensing, time-series analysis, and causal inference modeling to comprehensively evaluate the status, history, and structural quality of post-mining land cover in Virginia and Tennessee. To establish a baseline of current reclamation status, we first developed a novel classification framework integrating Sentinel-2 spectral data, CCDC phenological metrics, and 3DEP LiDAR structure to map seven land cover classes, with a specific focus on the invasive shrub Autumn olive (Elaeagnus umbellata). The classification achieved area-weighted overall accuracies between 74.3% and 76.8%, revealing that Autumn olive occupies approximately 9.8% of mined land in Virginia, peaking in prevalence on sites 14 years post-disturbance. Expanding this analysis temporally, we reconstructed a 37-year history (1984–2021) of land cover dynamics, documenting a region-wide decline in barren land from 60% to roughly 14–22%. However, succession trajectories diverged significantly by state; while Tennessee mines transitioned largely to coniferous and herbaceous cover, Virginia mines experienced a progressive expansion of Autumn olive to 14% of the landscape by 2021. Spatially explicit ANN-CA-MC simulations project autumn olive presence will persist into the near future, and that historical expansion is driven by transitions from herbaceous and shrub/scrub cover. Finally, to assess the quality of successfully reforested mines, we utilized causal random forests and sample balancing to quantify the effect of mining on forest structure compared to non-mine disturbances. We found that while canopy cover on mines converges with reference forests after 15–20 years, mine forests in Virginia remain significantly stunted, averaging 2.28 m shorter than forests recovering from non mining disturbances. Furthermore, standard spectral recovery metrics derived from Landsat may not full capture these structural deficits, highlighting the necessity of structural data for reclamation monitoring. Collectively, these findings demonstrate that while surface mines have successfully revegetated, they frequently fail to restore structural attributes and desired land cover composition, often diverting into stable, invasive-dominated states that require active management intervention
An Investigation of the Interplay Among CX3CR1, Immune Cells, and Gut Microbiota in Lupus-Associated Arthritis and Renal Disease using the MRL/lpr Mouse
Estaleen, Rana (Virginia Tech, 2026-01-09)
Discovering Viral Hosts, Mutations, and Diseases using Machine Learning
Antony, Blessy (Virginia Tech, 2026-01-09)
The discovery of a novel virus raises three important questions, namely, which host(s) can the virus infect, what mutations in the virus could affect its interaction with its hosts and enable a host-shift, and which diseases can the virus cause in humans. We propose novel machine learning (ML)-based solutions to these three different problems in computational virology. (i) We develop a viral protein language model for predicting the host infected by a virus, given only the sequence of one of its proteins. Our approach, 'Hierarchical Attention for Viral protEin-based host iNference (HAVEN)', includes a novel architecture comprising segmentation and hierarchical self-attention to tackle the challenges posed by long sequences. Pretrained on 1.2 million viral protein sequences, the model accepts any protein sequence of any virus and predicts its host. We integrate HAVEN with a prototype-based few-shot learning (FSL) classifier to generalize it to predict rare and unseen hosts, and hosts of unseen viruses. (ii) Structured datasets of known viral mutations and their effects are required to develop computational models that can predict potential detrimental changes in novel animal viruses. We leverage large language models (LLMs) to create these datasets from the results of experimental studies available as unstructured text in scientific literature. We design an open-ended task for 'scientific information extraction (SIE)' from publications and propose a unique two-step retrieval augmented generation (RAG) framework for the same. We curate a novel dataset of mutations in influenza A viral proteins. We use this dataset to benchmark our proposed approach, a wide range of LLMs, RAG-, and agent-based tools for SIE. (iii) Finally, we look at the effects of viral infections in humans. Specifically, we focus on the long-term effects of SARS-CoV-2 (or long COVID) wherein patients experience the persistence of COVID-19 symptoms for a long period of time after their initial SARS-CoV-2 infection. We propose an ML-based classification pipeline to predict the diagnosis of long COVID in COVID-19 patients using their electronic health records (EHRs) in the National COVID Cohort Collaborative, which is the largest collection of clinical data across the US. Using techniques to explain our models' prediction for each patient, we uncover many features that were correlated with long COVID. We also evaluate the impact of different data sources on our long COVID prediction models using a novel a cross-site analysis.
Perceptions of Leader Development Programming by College Students with Introverted Personalities
Martin, Perry Douglas (Virginia Tech, 2026-01-09)
This is a qualitative study on the perceptions of leadership development programming by students who identified as more introverted than their peers. The study examined the self-efficacy of these students towards leadership and the contributing factors to the achievement of their efficacy to be a leader. Conducted at a Research I, land-grant institution, the study consisted of interviews with students who identify as more introverted than their peers. Interviews allowed the researcher to examine their experiences and attitudes towards their own leadership development. The purpose of the study was to better understand the concept of leadership efficacy in the context of introverted student experiences. Findings from the study highlighted the importance of close relationships as a source of vicarious learning, verbal encouragement, and as a steadying influence on emotional well-being for introverted students developing as leaders. Students value teaching as an optimal model for leadership. As they navigated the rigors of serving in leadership roles in college, students looked to close relationships and regular practices of self-care to mitigate the impacts of stress on their energy. This study contributes to the body of knowledge on the understanding of personality and leadership development, specifically how self-efficacy is manifested in those with an introverted personality.
Investigating the Use of Physiological and Behavioral Signals to Facilitate Empathic Human-AI Interaction for Daily Stress Management
Dongre, Poorvesh (Virginia Tech, 2026-01-08)
This dissertation explores the design and evaluation of Empathic Large Language Models (EmLLMs) for general mental health support. EmLLMs use physiological and behavioral signals to infer users' mental states (affective and cognitive) and accordingly generate empathic messages as adaptive interventions. Three core research goals guided this work: (1) systematically reviewing state-of-the-art methods for stress and affect recognition with physiological signals and for designing physiologically adaptive systems, (2) developing and evaluating physiology-driven EmLLM prototypes that integrate stress detection with LLM-based dialogue for stress intervention, and (3) evaluating the performance and stability of multimodal LLMs using behavioral signals for emotion recognition and supportive message generation. Findings from the systematic review highlight that physiological signals provide valuable insights into stress and affect, and that systems with physiology-driven adaptation are effective at improving both user experiences and mental health interventions. Autoethnographic and pilot studies with graduate students on different prototypes of physiology-driven EmLLMs demonstrate promise for daily stress management, and expert evaluations provide further insights into refining the design of physiology-driven EmLLMs for real-world and clinical use. Performance and stability evaluations of multimodal LLMs show that multimodal behavioral inputs, including voice and facial features, enhance emotion recognition and reasoning. However, model behavior varies across modalities, underscoring the need for robust evaluation, customization strategies, and protective safeguards for mental health applications. Overall, this dissertation offers a systematic review, empirical insights, and design guidelines for developing empathic, engaging, and effective digital mental health systems.
Energy metabolism in skeletal muscle: mechanistic insights into mitochondrial function and growth
Yen, Con-Ning (Virginia Tech, 2026-01-08)
Skeletal muscle metabolism is critical to understanding the efficiency of muscle growth for the global meat industry. As such, there are many contributing factors that can influence muscle growth yet there are still basic molecular mechanisms that remain unexplored. Specifically, the involvement of mitochondria during skeletal muscle hypertrophy. Armed with this goal, we sought to target the fundamental aspects of how the mitochondria facilitate muscle growth in various livestock species. We showed that skeletal muscle mitochondrial abundance through mitochondrial DNA (mtDNA) and protein may not be sufficient to determine their functionality significance in muscle of livestock. Particularly, avian mitochondrial function is independent of absolute mtDNA and protein abundance. However, porcine and bovine muscle mitochondria abundance correlated with skeletal muscle type function. These findings confirm the importance of evaluating mitochondrial content and function to determine their overall contribution to muscle metabolism. To investigate the role of the mitochondria during muscle growth, we utilized beta-adrenergic agonists (BAA) fed pigs harboring the constitutively active adenosine monophosphate activated protein kinase mutation (AMPKγ3R200Q) that results in greater oxidative capacity in a habitually glycolytic skeletal muscle. We discovered BAA supplementation stimulates beta-1 adrenergic receptor gene expression and impacts mitochondrial respiration. Interestingly, BAA feeding had more effect on control pig muscle compared to that of pigs harboring the AMPKγ3R200Q mutation. This further suggests BAA-induced muscle hypertrophy is not as effective on muscles with increased oxidative capacity. To expand our knowledge on muscle hypertrophy, we assessed the impact of finishing feeding regimes on beef cattle muscle energy utilization. We showed that mitochondria from muscle of forage-maintained cattle have greater respiration compared to mitochondria of those carbohydrate-maintained cattle and glycolytic muscles. These data indicate that diet impacts skeletal muscle metabolism specifically in the ability of mitochondria to utilize long- and short-chain fatty acids. In addition to studying muscle hypertrophy in livestock, we generated knockout mouse models to assess the necessity of mitochondria in skeletal muscle during post-weaning growth. We found that reduced mtDNA content has a mild impact on mitochondrial protein expression and functionality, yet appears accumulative in its impact on skeletal muscle as reflected in final body weight and lean mass of knockout mice. Finally, we created a knockout mouse lacking a functional mitochondrial ATP synthase subunit beta gene and found that the ability of the mitochondria to generate ATP was not requisite for post-weaning growth of muscles expressing myosin light chain-1. In summary, mitochondria contribute significantly to overall skeletal muscle metabolism yet may not be required for optimal muscle growth prior to maturity. Further understanding of the molecular mechanisms necessary for optimal muscle growth is necessary to provide additional opportunities to improve efficiency of livestock growth.
Stories of the Wind
Soares Souza de Souza, Aline Regina (Virginia Tech, 2026-01-08)
Stories of the Wind is an audiovisual performance exploring various media to tell a story, integrating media at the intersection of visual arts and music, leveraged by technology. Different materials and technologies coexist as pieces of an audiovisual performance, with images, sound objects and interactive works. The production of this work was informed by artistic-scholarship, which involved the combination of aesthetic education and aesthetic experience with research and analysis in the process of artistic and academic creation. This project was meant to be exhibited in the Cube, at the Moss Arts Center, at Virginia Tech. Because of the Covid-19 pandemic, it was not possible to present the project in the space that it was created for, so a video adaptation was made to be submitted for the thesis defense. The video submitted as the thesis project pandemic adaptation can be seen through the following link: â€‹https://www.youtube.com/watch?v=dH8ce9KO41wandt=50s
Empathetic Educational Environments: Advancing Cultural Sensitivity of Trauma Through Storytelling in the Secondary-level English Classroom
Rose, Mackenzie Shannon (Virginia Tech, 2026-01-08)
This exploratory sequential mixed methods study addresses the critical gap in trauma-informed educational research by centering trauma survivors' voices as foundational knowledge for developing pedagogical interventions. First, through comprehensive evaluation of a teacher preparation program, classroom observations across diverse Virginia schools, educator professional development (n =7), and in-depth interviews trauma survivors (n =15), this research reveals significant deficits in trauma-informed practices within secondary education settings. The study introduces "misbehaving forms," alternative narrative structures that deliberately resist conventional academic constraints to accommodate the non-linear nature of trauma expression. This concept emerged with survivor narratives describing how traditional formats failed to capture their authentic experiences. These qualitative findings then informed the development of a storytelling intervention, which was implemented in three sections of a secondary English class and measured through an adapted Self-Determination Theory questionnaire assessing autonomy, competence, and relatedness. Two sections of the English class (n = 39) received the intervention, and one section remained as the control (n =16). With this questionnaire, quantitative results demonstrated statistically significant improvements in the intervention participants' student autonomy (p = 0.0309), competence (p = 0.0069), and creative expression (p < 0.001). The integration of qualitative and quantitative findings validates the effectiveness of survivor-informed pedagogical approaches while establishing a methodological framework for centering marginalized voices in educational research. The research challenges traditional academic hierarchies that exclude survivor wisdom while providing practical strategies for creating trauma-informed learning environments that support both academic achievement and emotional healing.
Targeting S1PR3 to mitigate flow-enhanced invasion in the glioblastoma tumor microenvironment
Howerton, Samantha Ann (Virginia Tech, 2026-01-08)
Glioblastoma is a devastating disease with few effective treatments, in part owed to the dynamic cellular and biophysical factors that influence tumor progression and therapy response. Emerging evidence has implicated pathological interstitial fluid flow, created by high intratumoral pressure relative to the healthy parenchyma, in enhancing cancer invasion. Multiple targetable molecular pathways have been identified that drive this response, but the specific pathways employed by invasive cells differs between patient glioma cell lines. To this end, we sought to identify additional therapeutic candidates mediating flow-enhanced invasion. Our previous work established a role for the G-protein coupled receptor S1PR3 in enhancing invasion under flow. Interestingly, we found this response to be mediated by the brain parenchymal cells, astrocytes and microglia. In this work, we demonstrate clinical relevance for S1PR3 as both a biomarker and therapeutic target with efficacy across a heterogeneous patient cohort. To inform therapeutic development, we investigate the intercellular mechanisms involved in S1PR3-driven invasion. We find that S1PR3 targeting significantly alters flow dynamics in vivo. We connect astrocytic S1PR3 to flow response, finding correlations with flow in both tumor-bearing and tumor-naïve settings, suggesting redundancy across neuropathologies. We build evidence that astrocytic S1P and S1PR3 mediates the response to fluid shear stress and we imply roles for S1P and flow-sensing. This work has exciting implications suggesting a dual role for S1PR3 in flow-regulation and flow-response, thus it may be a doubly effective target for minimizing flow-enhanced glioma invasion.
Aerodynamic Enhancement and Reduced Order Modeling of Vertical Axis Wind Turbines
Shanab, Belal (Virginia Tech, 2026-01-08)
Vertical-axis wind turbines (VAWTs) are recognized as a viable solution for wind energy harvesting. This study discusses VAWT performance in different aspects. First, using computational fluid dynamics (CFD) simulations, the impact of various deflector angles as an auxiliary augmentation on turbine efficiency is examined. Specifically, the deflector orientation angle effect on the dual-rotor straight blade vertical-axis wind turbine (DR-SBVAWT) performance is conducted. Two-dimensional transient simulations were performed for this parametric study. The results demonstrate that deflector implementation boosts the DRSBVAWT self starting capabilities and enhances overall performance. With a vertical deflector (β = 0◦from the y-axis) yields the best performance, providing the highest efficiency and power output. In contrast, a horizontal deflector (angle of β = 90◦ counterclockwise from the y-axis) shows minimal impact on the turbine's performance, suggesting that further angle variations do not significantly enhance the system. Moreover, sensitivity analysis was performed to evaluate the impact of small changes in the deflector orientation that shows β = 0◦ holds the best orientation and small angle variations in deflector orientation show minimal impact on the overall performance of the turbine. In steady-state conditions, the vertical deflector angle increases the tip speed ratio (TSR) of the DR-SBVAWT performance by 11.5% compared to a conventional DR-SBVAWT without the deflector. Additionally, this vertical configuration achieved a 30.15% increase in efficiency at TSR = 2.5, showing its effectiveness in improving overall aerodynamic performance. This parametric study overall provides valuable insights into the optimal deflector angle configuration as an auxiliary augmentation system for dual-rotor vertical axis wind turbines, contributing to the design optimization and improved performance of wind energy systems. Secondly, utilizing the same method of a 2D transient unsteady Reynolds-averaged Navier– Stokes (URANS) numerical simulations, different clustering scenarios of DR-SBVAWT for farm design are studied. Twelve clustering configurations include vertically aligned pairs and staggered clusters of three turbines at different inter-turbine distances, to evaluate their impact on power capture and land usage for the DR-SBVAWT, are investigated. Performance indices, namely, total power coefficient and improvement relative to standalone turbines, are analyzed. Wake effects are qualitatively discussed through detailed velocity contour plots of the wind field. Results show that a DR-SBVAWT turbine arrangement can enhance wind farm performance by approximately 25% for two turbines and about 20% for three staggered turbines, with required spacing of 1.5D and 2.5-3D, respectively. The study overall provides critical insights into the optimal placement and configuration of DR-SBVAWTs for maximizing energy output while minimizing land usage, offering guidance for the design of more efficient VAWT farms. Furthermore, it offers guidance on mitigating destructive interference from upstream turbines, enabling the optimization of multi-turbine layouts so that each unit operates under stable flow conditions. Finally, a reduced order model (ROM) is investigated for VAWT. A novel and robust CFDROM framework is built to evaluate the complex flow field behaviour of a single rotor 3-blade VAWT. A data-driven framework was developed following high-fidelity transient simulations conducted using the URANS simulations within a finite volume method (FVM) framework in ANSYS Fluent. A snapshot-based proper orthogonal decomposition (POD) reduced-order model was developed. Additionally, a domain decomposition strategy was implemented to analyze the flow behavior across the interface between subdomains (the rotor domain and its surrounding domain) of VAWT. To ensure accurate enforcement of interface continuity conditions, the coupling between the inner and outer domains was achieved through a tunable proportional controller (computational gain). Results reveal that domain decomposition approach coupled with ROMs enhances the robustness of the computational cost while maintaining acceptable accuracy. By adding the computational controller, it helps to improve the coupling of the two ROMs, ensuring more efficient integration with fewer mismatches at the interface between the decomposed domains. The data-driven novel framework provides a critical stepping point to extending research to include more complexity of flow dynamics of VAWTs. This proposed framework proves a speedup on the order of one for 300 snapshots, or on about order of four for complete transient simulations (20 seconds) of fixed VAWT. Overall, given the greater viability and applicability of VAWTs in different setups compared to other turbine types, these three detailed studies of dual-rotor design, farm design, and ROM for VAWT, offer a comprehensive framework for understanding VAWT aerodynamics and provide more efficient, high-fidelity design and analysis that benefits in varied wind environments.
Assessing Extension Engagement and Collaboration Through the Virginia Cooperative Extension Situation Analysis Process
Johnson, Lonnie L. Jr. (Virginia Tech, 2026-01-08)
Engagement and collaboration among Extension personnel and community stakeholders are critical for maintaining the relevance, effectiveness, and sustainability of Cooperative Extension programs. This study examined internal and external perceptions of collaboration during the Virginia Cooperative Extension (VCE) Situation Analysis process, which is a formal, statewide assessment conducted every five years to identify and prioritize local community issues. The purpose of this study is to gain additional insight into the perceptions of collaboration between and among employees and stakeholders who participated in the VCE Situation Analysis process and to better understand areas for professional development for employees and stakeholders regarding this process. The objectives of this study were to: (1) describe the perceived strength of collaboration from internal and external perspectives, (2) compare these perceptions across groups, (3) identify differences or similarities in perceived collaborative strength, and (4) determine areas for employee development to enhance future engagement. A modified Wilder Collaboration Factors Inventory (WCFI), excluding the Environment and Resource domains, was administered via Qualtrics to 420 identified participants, yielding 167 responses (39.7% response rate). Quantitative data were analyzed across four collaboration domains: Membership Characteristics, Process/Structure, Communication, and Purpose. Mean ratings across all respondents who completed the survey (n = 102) indicated moderate to strong collaboration (M = 3.71 on a 5-point Likert scale), with the Purpose domain receiving the highest mean (M = 3.78) and Process domain the lowest (M = 3.64). External stakeholders consistently rated collaboration higher (M = 4.10) than internal participants (M = 3.63). The highest-rated indicator was "I have a lot of respect for other people in this collaboration" (M = 4.20), while the lowest was "All of the organizations that we needed to be members of this collaborative group became members" (M = 3.12). Qualitative data from open-ended survey responses (N = 50) and 16 follow-up interviews identified recurring themes of community connection, issue identification, and the need for enhanced communication and engagement strategies. Participants emphasized that stakeholder involvement improves the accuracy of community needs assessments, strengthens trust, and supports the development of relevant, high-impact programs. Key skills identified for effective collaboration included soft skills (listening, communication, facilitation, leadership) and technical competencies (data analysis, process knowledge, time management). Findings suggest that while VCE demonstrates strong collaborative intent and mutual respect, gaps remain in stakeholder inclusion and communication consistency. Improving stakeholder engagement through targeted employee development, intentional outreach, and clearer process communication could strengthen future Situation Analyses. Overall, the study underscores that effective collaboration not only enhances Extension's local relevance but also reinforces accountability and support at federal, state, and local levels.
Engineering Tumor-Targeting Bacteria and Characterizing Their Interactions with Tumor Cells in Therapy-Resistant Breast Cancers
Saha, Binita (Virginia Tech, 2026-01-08)
Breast cancer (BC) accounts for one-third of malignancies among women in 157 countries and ~15% mortality among diagnosed cases, a burden projected to reach 1.1 million deaths per year by 2050. Five-year survival drops from >90% in early-stage BC to ~32% for therapy-resistant subtypes such as triple-negative (TNBC), hormone-receptor–variable or resistant ER+, and Luminal B tumors. For these high-risk subtypes, molecular heterogeneity undermines targeted therapies, and clinical management still relies on maximum tolerable doses of systemic chemotherapy, causing severe dose-limiting toxicities. Moreover, the dense collagen-rich extracellular matrix (ECM) of solid tumors restricts intratumoral drug transport, motivating strategies that function across BC subtypes, overcome ECM barriers, and inform lowered clinical dosing. Bacteria-based cancer therapy (BBCT) with cancer-selective bacteria combines motility, self-replication, and on-board biosynthesis with the programmability of synthetic biology, enabling local release of therapeutic factors within the tumor microenvironment. Attenuated Salmonella Typhimurium VNP20009 (ST) exhibits ~10³-10⁴-fold tumor selectivity with respect to liver and spleen. It has a favorable clinical safety profile but remains inefficacious due to poor colonization. Our lab previously showed that ECM-targeting ST with collagenase secretion improved tumor penetration without gross collagen disruption, but at the cost of reduced bacterial fitness and motility. In this dissertation, we hypothesized that a fitness-restored, ECM-targeting ST enhances bacterial intratumoral transport and colonization, as well as chemotherapy penetration. We evaluated the engineered strains in perfused 3D tumor models that represent in vivo intratumor transport properties and investigated cancer cells-neutrophil interactions in presence of bacterial factors. First, we developed a high-motility, fitness-improved collagenase-expressing strain (HM-CEST ΔydcP) that preserves 100% motility under sub-cytotoxic chemotherapy. This strain improves intratumoral transport, and reduces spheroid viability and tumor migration relative to chemotherapy alone while maintaining tumor specificity and safety in preclinical murine models in vivo. Second, we validated a perfusion-enabled microfluidic spheroid platform that supports at least 14-day culture of murine TNBC and ER+ spheroids, enabling long-term BBCT screening. We demonstrated that perfused spheroids require lower chemotherapy doses than static cultures. Third, we biophysically characterized the crosstalk between BBCT, neutrophils, and Luminal B BC cells, demonstrating that neutrophils in the presence of bacteria supernatant suppress cancer cell growth, viability, and migration. Collectively, this work delivers a fitness-restored ECM-targeting Salmonella chassis, validates a perfusion-enabled 3D tumor-spheroid microphysiological platform, and develops a quantitative framework for neutrophil–bacteria–cancer cell interactions, contributing to the design pipeline for next-generation BBCT and supporting the long-term goal of safer, more affordable, lower-dose, and more broadly applicable therapies for therapy-resistant breast cancers.
Novel Insights into Lead Corrosion Control: Safely Changing Source Waters, Effective Use of Zinc Phosphate Inhibitors, Electrochemical Reversal of Solder, and Significance of Lead-Polyphosphate Complexation
Mazzola, Frank Anthony (Virginia Tech, 2026-01-07)
Recent high profile lead contamination events arising from lead solder corrosion have drawn attention to our alarmingly inadequate understanding of corrosion control chemistry. This dissertation provides novel contributions to an improved understanding of lead corrosion control in four sequential chapters that: 1) develop a new framework to proactively evaluate the impact of source water changes on lead solder corrosion, 2) demonstrate that in certain corrosive waters zinc orthophosphate can dramatically reduce lead contamination from solder at pHs around 7.5-9.5, but is ineffective in relatively non-corrosive waters or at pHs outside this range, 3) reveal that prolonged exposure to free chlorine can dramatically reduce lead contamination from solder by causing "electrochemical reversal," protecting solder from galvanic corrosion, and 4) apply a method for measuring problematic lead complexation by polyphosphate, which reveals a significant health concern in some water chemistries but insignificant problems in others. Chapter 2 describes how water utilities are increasingly changing source waters in response to groundwater contamination, sustainability efforts, new regulations, and consolidation. We worked with three communities who unexpectedly experienced very severe lead contamination from lead-tin solder following a seemingly innocuous change in source water. We demonstrate how existing frameworks did not anticipate lead solder corrosion problems because: 1) prolonged exposure to non-corrosive water can preserve lead solder in relatively pristine condition, which is more susceptible to severe corrosion following changes in water chemistry, and 2) prolonged exposure to corrosive water can be self-limiting because all exposed lead in copper pipes could have corroded away within a period of decades, counter-intuitively producing low present day lead contamination in a highly corrosive water. A failure to understand the implications of these scenarios resulted in erroneous logic predicting that a source water change would be safe when it was not. We develop a proactive bench scale testing protocol and improved framework that can allow utilities to avoid similar mistakes in the future. Chapter 3 addresses divergent results regarding the effectiveness of zinc orthophosphate corrosion inhibitor in controlling lead solder corrosion; that is, some research reported remarkable advantages compared to adjusting pH/alkalinity or dosing orthophosphate alone, whereas other research found no benefits. We examined relative performance of zinc orthophosphate at a wide range of pHs (6.5, 7.5, 8.5, 9.5, 10.5) in a low alkalinity, low conductivity water with corrosivity increased by addition of extra chloride or nitrate, and demonstrate that the results depend on water chemistry. Specifically, zinc orthophosphate has little benefit in water with low chloride or nitrate, in which corrosivity is low, at any pH tested. But at higher corrosivity zinc orthophosphate significantly reduced lead release at pH 7.5-9.5, including 40X reductions in lead release at pH 8.5. Yet no relative advantage occurred at pH 6.5 and pH 10.5. Results are consistent with passivation from zinc phosphate precipitation from pH 7.5 to 9.5 as predicted by equilibrium modeling and supported by scale analysis. Similar methods can be used to predict relative performance of inhibitors for controlling lead solder corrosion in other source waters and determine if these results are generalizable. Chapter 4 explores a 40-year-old mystery with profound present-day implications. A study in Portland, Oregon had revealed that free chlorine had orders of magnitude less lead contamination when compared to chloramine, but the result was discounted because chlorine was always believed to be more corrosive than chloramine. Here, we resolve the controversy by demonstrating "electrochemical reversal" of the copper-solder galvanic couple in synthesized water similar to Portland's. That is, short-term exposure to chlorine did not reduce lead contamination, but prolonged exposure to free chlorine caused the normally anodic solder to become cathodic via formation of a Pb(IV) scale. This eventually caused free-chlorine treated water to have 10-100X less lead contamination than chloramine-treated water after several months of exposure. This discovery has major implications for water treatment and can also help explain why some waters dosed with free chlorine have anomalously low lead contamination. Chapter 5 examines fears that utilities adding polyphosphates to drinking water to sequester iron and manganese, or to prevent calcium carbonate scaling, will invariably suffer from higher lead solubility and a higher risk of lead contamination. Here, we develop a simple semi-quantitative method using cation exchange resins to evaluate the significance of lead-polyphosphate complexation. Applying this test in a range of waters, we reveal the dependency of lead-polyphosphate complexation on water hardness, polyphosphate dose, and the ratio of orthophosphate to polyphosphate. This approach can be used to predict the impact of a given polyphosphate product and dose on lead complexation, allowing utilities to predict the magnitude of the problem for a given water in about a day. We find a strong linear relationship between polyphosphate dose and complexed lead. Additionally, when >200 mg/L as CaCO3 of calcium is present, >3X less lead-polyphosphate complexation can occur versus when no calcium is present. Our work also demonstrates that testing artifacts probably caused past research to overestimate the danger of the lead complexation problem. This work has significant implications for public health and is timely as water lead contamination is increasingly scrutinized. The simple tests and corrosion control methods described herein can help drinking water utilities make informed water treatment decisions to reduce lead contamination and protect public health.
Enhancing side-channel analysis through measurement, and high-power IEMI generation
Singh, Alok Kumar (Virginia Tech, 2026-01-07)
In today's interconnected world, the use of hardware security modules (HSMs) or trusted platform modules (TPMs) has been growing rapidly. These devices are the foundation of many security measures, using cryptographic algorithms to ensure the confidentiality and integrity of sensitive data. For example, an HSM in the vehicle's electronic control units (ECU) safeguards vehicle communications and functional control systems using cryptography. However, these devices are not immune to attacks, as an adversary can gain easy physical access (or be in close vicinity) to the device or communication medium. One such attack is side-channel analysis (SCA). This work proposes an effective methodology to launch power SCA and increase the efficiency of the attack by improving the measurements. The research examines heuristics related to measurement parameters, investigate ways to optimize the parameters, determine their effects empirically, and provide a theoretical analysis to support the findings. This work introduces a novel, measurement-focused methodology that is attack-agnostic, leveraging multi-sensor fusion with a Kalman filter to enhance SCA data resolution and significantly reduce the number of measurements needed for successful attacks. We propose and realize a low-cost, low-noise, multi-sensor measurement board to demonstrate the effectiveness of our approach. The board enables the independent but coupled measurement of both a device's power consumption and the associated electromagnetic field it produces, which we combine with a Kalman filter to improve the accuracy of the power measurement. This enhanced data quality can significantly boost the efficiency of SCA, independent of the chosen attack method(s). The second phase of this research investigates intentional electromagnetic interference (IEMI), a wireless attack where an adversary uses an electromagnetic field in close proximity to induce a specific secondary effect on a target device. Unlike typical cyberattacks that exploit software vulnerabilities, this attack bypass conventional cybersecurity defenses by targeting the hardware layer directly with limited or zero physical access to the target device. The research focuses on the hardware architecture and design of two distinct amplifier types: one capable of operating across a wide range of frequencies, and a second that functions as a high-power single-tone amplifier capable of sourcing power to radiators in the kilowatts range. This work demonstrates the effectiveness of the proposed hardware through two distinct applications: wireless vehicle fingerprinting and a novel "wireless spiking" technique on smart locks, where an attacker wirelessly bypasses standard security measures to lock or unlock the device.
Towards Secure and Resilient Machine Learning Systems
Sun, Shihua (Virginia Tech, 2026-01-07)
Over the past decade, Machine Learning (ML) technologies have undergone revolutionary advancements, extending beyond traditional domains such as computer vision (CV) and natural language processing (NLP). One of the most significant breakthroughs is the development of transformer models, which leverage the attention mechanism to achieve state-of-the-art performance across various tasks. Transformers serve as the foundation for commercial large language models (LLMs), such as GPT and Claude, driving progress in natural language understanding and generation. Beyond natural language, transformer architectures have been successfully adapted to source code analysis by pretraining and fine-tuning models on large corpora of programming languages. In parallel, the emergence of Vision Transformers (ViTs) has demonstrated exceptional performance in CV applications, further challenging the dominance of convolutional neural networks (CNNs). Another transformative advancement is Federated Learning (FL), a decentralized learning paradigm that preserves data privacy while enabling collaborative model training across distributed clients. Given its advantages in privacy-sensitive domains, FL provides a compelling foundation for cybersecurity applications, particularly for enhancing Intrusion Detection Systems (IDSs) in IoT networks. Its decentralized nature makes it well-suited for Internet of Things (IoT) ecosystems, where data is generated across diverse devices, offering an effective solution for both privacy protection and robust threat detection. However, integrating ML models into real-world applications exposes them to adversarial threats. These include poisoning attacks in the training phase and evasion attacks during inference, both of which compromise model reliability and accuracy. To enhance the robustness of ML models, this dissertation presents a series of studies that (1) strengthen the resilience of ViTs against evasion attacks, (2) investigate the vulnerabilities of FL to advanced poisoning attacks, (3) develop FL-based IDSs for IoT networks that effectively address performance degradation caused by data heterogeneity, and (4) analyze the robustness of transformer models pretrained on programming languages against code-based evasion attacks and propose effective strategies to strengthen their defenses. Collectively, these contributions aim to improve the security, adaptability, and effectiveness of ML models in real-world deployments.
A framework for Improving Hydrologic and Water Quality Prediction in Urbanized Watersheds through Stakeholder Co-Design and Multi-Model Integration
Shah, Vishwa Satishbhai (Virginia Tech, 2026-01-07)
Urban watersheds present a unique modeling challenge due to the complex interplay between natural and hydrologic processes, engineered infrastructure, and the diverse decision-making by multiple stakeholder groups. These interactions span multiple spatial and temporal scales, making it difficult for any single modelling approach to represent the system's full complexity or adequately address diverse stakeholder needs. Many existing modeling frameworks fail to align with stakeholder decision processes, reducing their relevance for applied watershed management. As the first objective, this dissertation introduces a stakeholder driven collaborative design process for developing the Occoquan Watershed Modeling Framework (OWMF), a multi-model, co-designed watershed modeling framework for simulating water quantity and quality applied within the Occoquan Watershed in Northern Virginia, USA. The framework represents a novel advancements in watershed modeling by addressing persistent design limitations in existing approaches by: (1) supporting multi-functional design objectives across hydrologic and water-quality domains, (2) embedding stakeholder priorities from the outset through an iterative, user-centered co-design process, (3) integrating scientifically rigorous, high-fidelity models, and (4) applying competency-based evaluation criteria to quantify performance, feasibility, and decision relevance. These design principles were operationalized through a structured, iterative co-design process that translated stakeholder priorities and model competencies into an implementable framework incorporating models such as GR4J-CemaNeige, SWAT, WAMRF and StormWise. The model selection was further validated by site-specific prototyping and performance evaluation. Following this, the finalized framework development plan was obtained through the co-design process. The analysis presented here provides a structured methodology to build stakeholder-driven, multi-model frameworks that can predict the short- and long-term impacts of natural and anthropogenic drivers that influence watershed resilience. The conclusions aim to bridge the gap between hydrologic modeling and watershed management, enabling a transparent, adaptive and transferable approach for enhancing watershed resilience. Understanding how future land use – land cover (LULC) and climate change (CC) can alter watershed hydrology and water quality is critical for effective long-term watershed management and planning. With the second objective, this dissertation incorporated a multi-model approach for improving the watershed-scale impact assessment under rapid urbanization and climate change. First, high-resolution LULC and CC projections were developed for the year 2040. Second, the watershed-scale dynamics under baseline (present) and future (2040) scenarios were simulated using three models: SWAT, HSPF, and WARMF. Third, an inter-model comparison was conducted that related the differences in model architecture, spatial discretization, process characterization and calibration strategy to watershed responses under future scenarios. The differences in simulated streamflow, pollutant loads (e.g., nitrogen, phosphorus), and sediment loads were quantified across the three models and three future scenarios. Despite using the same forcing, the three models produced different magnitudes of change in streamflow, sediments and nutrient loading, reflecting the impact of model structure in affecting processes such as simulated runoff generation, sediment detachment, subsurface flow partitioning, phosphorous transport and nitrogen cycling. Moreover, the simulation timestep (hourly vs daily), calibration timestep (hourly vs daily vs monthly) and input data resolution directly impacted the sensitivity of these models to LULC change and climate variability. The inter-model comparison concluded that in addition to the model structure, the calibration methodology impacted how the models projected into the future. Whether the calibration was biased or unbiased towards extremes and which objective functions (streamflow, ET, nutrients etc.) were chosen for calibrating the baseline models had a profound impact on predicting the future watershed responses across the three models. The study showed that multi-model assessments should be the standard methodology for improving confidence in future watershed-scale hydrologic and water quality predictions under the influence of future variability in LULC and climate. Seasonally shifting hydro-meteorological conditions can introduce substantial variability in watershed response, yet majority of rainfall–runoff models often rely on fixed parameter sets that do not adjust to these changes. The third objective incorporated a multi-pronged approach for improving seasonality representation by improving model parameterization and coupling it with data-driven modeling of hydrologic systems. This study leveraged both these approaches for representing seasonality in hydrologic models for improved streamflow prediction. Using the GR4J-CemaNeige model for the Occoquan Watershed in Northern Virginia, this study tested the application of a predefined four-season parameterization, followed by univariate and multivariate clustering to identify data-driven hydro-climatic patterns. Insights from these analyses informed the development of a hybrid dynamic parameterization in which model parameters varied continuously with time and varied with respect to local-scale potential evapotranspiration observations. Results showed that traditional four-season parameterization improved hydrologic performance only when seasonal boundaries coincided with actual hydrometeorological behavior. The univariate clustering analysis showed that temperature and evapotranspiration followed repeatable annual cycles, whereas precipitation and streamflow displayed irregular and highly variable seasonal behavior, including transitional months without consistent cluster identity. The multivariate clustering further demonstrated that the combined hydro-climatic variables did not align reliably with fixed patterns, reflecting the irregular timing of hydrologic conditions in the study basins. The dynamic formulation generated continuously evolving parameter trajectories and produced more consistent performance across evaluation periods. Collectively, the stepwise progression, from seasonal calibration to clustering-based diagnostics and dynamic parameterization provided a systematic framework for diagnosing seasonal hydrologic behavior and enhancing the temporal adaptability of conceptual hydrologic models.
Interpretability and Debugging for Distributed Privacy Preserving Machine Learning
Gill, Waris (Virginia Tech, 2026-01-07)
Machine learning systems increasingly rely on privacy-preserving distributed training to leverage sensitive data across multiple organizations without centralization. Federated Learning (FL), a distributed privacy-preserving machine learning paradigm, enables hospitals, devices, and enterprises to collaboratively train models without accessing raw client data (e.g., Siri, Alexa, and healthcare applications). Centralized machine learning benefits from rich debugging and interpretability techniques enabled by transparent access to training data. However, FL removes this transparency, rendering traditional techniques ineffective and making debugging and interpretability a challenging open problem. This thesis addresses this challenge by asking: How can we design automated debugging and interpretability methods for federated learning that effectively localize faults and attribute global model predictions without degrading performance or violating FL's core privacy principles? The central insight is that effective debugging and interpretability can be achieved by analyzing model parameters, activations, and gradients-information already shared or derivable in standard FL protocols (e.g., FedAvg). We present three contributions. First, towards fault localization, we redesign traditional differential testing to operate on neuron activations produced by auto-generated inputs, exploiting the fact that faulty clients produce models with divergent activations. Second, we introduce neuron provenance, which decouples data-influence tracking from data access. It identifies influential neurons via gradient-based weighting and decomposes them to client-specific origins, yielding ranked lists of responsible clients across CNNs and Transformers. Third, we extend neuron provenance to federated LLMs, where autoregressive generation and billion-parameter scale make naive tracking infeasible. It introduces token-level provenance at targeted transformer layers, achieving high attribution accuracy across multiple LLM architectures. In each case, the solution operates entirely on information available at the aggregator, requiring no client-side instrumentation. Collectively, these contributions culminate in practical tools that integrate seamlessly with existing distributed ML workflows, enabling real-time debugging and transparent model insights for both classification and LLMs in FL.
Graph-Based Computational Approaches for Modeling Viral Evolution
Das, Badhan (Virginia Tech, 2026-01-07)
Modeling viral evolution is essential for understanding how pathogens adapt, spread, and generate new variants of concern. Yet, it remains challenging due to high mutation rates, minimal sequence divergence, and the scale of modern genomic data. Most phylogenetic trees enforce a strictly bifurcating structure that struggles to represent recurrent mutations, recombination, convergent evolution, and intra-host diversity. In contrast, quasispecies theory describes viral populations as clouds of closely related mutants evolving within a high-dimensional sequence space, where evolutionary relationships are more naturally captured by graphs than trees. In this dissertation, I develop a sequence of graph-centered frameworks that integrate viral fitness, mutational distance, and mutational dynamics to model viral evolution from algorithmic and data-driven perspectives. First, ViraFit introduces a proof-of-concept model that couples epidemiological spread on contact networks with evolutionary dynamics on fitness landscapes, demonstrating how mutation, selection, and network structure jointly shape adaptive trajectories. Second, the Variant Evolution Graph (VEG) provides a scalable graph-based representation of SARS-CoV-2 evolution derived from mutational distances, allowing multiple ancestral relationships and capturing virus-specific evolutionary patterns that are difficult to represent with phylogenetic trees. A derived Disease Transmission Network further supports inference of likely transmission pathways and superspreaders. Finally, the Ancestor-Joining algorithm extends this representation into a predictive framework, Mutation Learning Graph (MLG), by inferring intermediate ancestral variants and enabling graph neural network–based lineage classification and mutational link prediction across geographically diverse SARS-CoV-2 cohorts. Together, ViraFit, VEG, and MLG form a unified methodological progression that links mechanistic modeling, evolutionary reconstruction, and predictive graph learning, providing a scalable, mutation-centric view of viral evolution that complements traditional phylogenetic approaches and supports future variant forecasting.
Distributionally Ambiguous Stackelberg Combinatorial Games for Submodular Optimization and Camera View-Frame Placement
Park, Seonghun (Virginia Tech, 2026-01-06)
This dissertation develops exact solution methodologies for Stackelberg zero-sum games, which model sequential decision-making between an attacker and a defender. Our work specifically addresses challenging settings where the defender's recourse is a complex com- binatorial optimization problem and the attacker faces uncertainty and distributional am- biguity. We analyze these games through two complementary frameworks. Distributionally Robust Optimization (DRO) framework provides a risk-averse attacker with robust attack- ing strategy, while offering defender insights into the most probable threats. In contrast, Distributionally Risk-Receptive (DRR) frameworks provides high-impact strategy for a risk- receptive attacker, thereby serving as a powerful tool for the defender's vulnerability analysis by exposing the system's most critical weakness. This dissertation makes three primary contributions, each developing novel decomposition methods based on structural insights. First, we introduce and solve a Stackelberg game, where the defender's problem is the camera view-frame placement problem. We address this setting under a DRO framework to explicitly model uncertainty in attack success, incomplete information, and the adversary's varying levels of risk-appetite. For this setting, we develop a cutting-plane-based algorithm that leverages a key geometric property: an optimal placement under one attack remains a feasible recourse under any other, to derive a new class of valid inequalities. Since our algorithm repeatedly solves the defender's problem, placing p camera view frames to maximize the coverage, we also contribute efficient exact methods for p = 1 and novel heuristics for p ≥ 2, validated through simulation experiments of finding a hidden object. Second, we solve the game when the defender's objective is maximizing k-submodular func- tion, under both DRO and DRR frameworks. To solve problem, we derive valid inequalities from the diminishing property of k-submodular function, and strengthen them further by imposing an ordering over elements in the defender's solution sets. The optimal values from these dual frameworks offer a confidence interval-like range for the defender's expected out- come, where the DRO solution provides robust attack strategies and the DRR solution iden- tifies critical data vulnerabilities. We demonstrate effectiveness of our frameworks through computational experiments on instances of feature selection and sensor placement problems, using Wisconsin breast cancer data and synthetic data, respectively. Third, we extend the strategic scope to a three-stage Defender-Attacker-Defender (DAD) model with fortification, where the defender's final recourse is the maximization of a sub- modular function. To solve this game, where the standard attacker-defender interdiction game appears as a subproblem, we derive another class of valid inequalities that are con- structed for an arbitrary fortification strategy by leveraging the diminishing return property of the defender's objective function. Empirical validation on real-world datasets with pre- dictive models (e.g., Support Vector Classifiers, logistic regression) confirms the practical impact of our frameworks.

Browse

Recent Submissions