Computer Sciences Commons^™

Leveraging Large Language Models For Knowledge-Free Weak Supervision In Clinical Natural Language Processing, Enshuo Hsu, Kirk Roberts

Faculty, Staff and Student Publications

The performance of deep learning-based natural language processing systems is based on large amounts of labeled training data which, in the clinical domain, are not easily available or affordable. Weak supervision and in-context learning offer partial solutions to this issue, particularly using large language models (LLMs), but their performance still trails traditional supervised methods with moderate amounts of gold-standard data. In particular, inferencing with LLMs is computationally heavy. We propose an approach leveraging fine-tuning LLMs and weak supervision with virtually no domain knowledge that still achieves consistently dominant performance. Using a prompt-based approach, the LLM is used to generate weakly-labeled …

Llms In Network Intrusion Detection – A Comprehensive Analysis, Sudharshan Balaji

USF Tampa Graduate Theses and Dissertations

Network Intrusion Detection Systems (NIDS) play a critical role in identifying and mitigating malicious activities within computer networks. With the rapid evolution of natural language processing (NLP), Large Language Models (LLMs) have emerged as transformative tools across various domains. LLMs, such as OpenAI’s GPT series and Meta’s LLaMA models,have demonstrated remarkable performance in tasks like language generation, reasoning, and classification. Their ability to understand and process vast amounts of data has enabled groundbreaking advancements in areas like healthcare, finance, and cybersecurity. Recent trends highlight their potential to handle unstructured data, perform complex reasoning, and adapt to a wide range of …

Rapid Prediction Of Coastal Flooding With Deep Neural Networks, Ali Shahabi, Navid Tahvildari

Graduate Student Government Association Research Conference

With the increasing impact of climate change and relative sea level rise, low-lying coastal communities face growing risks from extreme storm tides and recurrent nuisance flooding. Thus, timely and reliable predictions of coastal water levels are critical to resilience in vulnerable coastal areas. Over the past decade, enormous efforts have been made to utilize machine learning (ML) based data-driven models for the emulation and prediction of storm tides. However, flood advisory systems still rely on running computationally demanding real-time hydrodynamic models. because developing highly reliable ML-based models suitable for real-time forecasting and capable of capturing any surge levels is challenging. …

Bridging Expert Knowledge With Deep Learning Techniques For Just-In-Time Defect Prediction, Xin Zhou, Donggyun Han, David Lo

Research Collection School Of Computing and Information Systems

Just-In-Time (JIT) defect prediction aims to automatically predict whether a commit is defective or not, and has been widely studied in recent years. In general, most studies can be classified into two categories: 1) simple models using traditional machine learning classifiers with hand-crafted features, and 2) complex models using deep learning techniques to automatically extract features from commit contents. Hand-crafted features used by simple models are based on expert knowledge but may not fully represent the semantic meaning of the commits. On the other hand, deep learning-based features used by complex models represent the semantic meaning of commits but may …

Deep Learning-Based Auto-Segmentation For Liver Yttrium-90 Selective Internal Radiation Therapy, Jun Li, Wookjin Choi, Rani Anne

Department of Radiation Oncology Faculty Papers

The aim was to evaluate a deep learning-based auto-segmentation method for liver delineation in Y-90 selective internal radiation therapy (SIRT). A deep learning (DL)-based liver segmentation model using the U-Net3D architecture was built. Auto-segmentation of the liver was tested in CT images of SIRT patients. DL auto-segmented liver contours were evaluated against physician manually-delineated contours. Dice similarity coefficient (DSC) and mean distance to agreement (MDA) were calculated. The DL-model-generated contours were compared with the contours generated using an Atlas-based method. Ratio of volume (RV, the ratio of DL-model auto-segmented liver volume to manually-delineated liver volume), and ratio of activity (RA, …

Semi-Supervised Multimodal Multi-Instance Learning For Aortic Stenosis Diagnosis, Zhe Huang, Xiaowei Yu, Benjamin S. Wessler, Michael C. Hughes

Computer Science Faculty Research & Creative Works

Automated interpretation of ultrasound imaging of the heart (echocardiograms) could improve the detection and treatment of aortic stenosis (AS), a deadly heart disease. However, existing deep learning pipelines for assessing AS from echocardiograms have two key limitations. First, most methods rely on limited 2D cineloops, thereby ignoring widely available Spectral Doppler imaging that contains important complementary information about pressure gradients and blood flow abnormalities associated with AS. Second, obtaining labeled data is difficult. There are often far more unlabeled echocardiogram recordings available, but these remain underutilized by existing methods. To overcome these limitations, we introduce Semi-supervised Multimodal Multiple-Instance Learning (SMMIL), …

Leveraging Distributed Semantics From Deep Learning Architectures For Literature-Based Discovery, Clint A. Cuffy

Theses and Dissertations

Literature-based discovery (LBD) is a scientific process that introduces methods to automatically identify novel insights between non-interacting sets of literature. To date, numerous statistical and machine learning-based methods have been applied in the biomedical domain to find treatments for diseases such as Raynaud's disease, Parkinson's disease, and Multiple Sclerosis. However, the lack of standardized practices and creation of bespoke methodologies produces a scenario where the adoption of LBD remains challenging in real-world systems. Our work addresses these concerns through the improvement of five critical areas: 1) error propagation within LBD's a priori dependent tasks, 2) exploring the integration of modern …

A Deep Learning Approach For Identifying Malicious Activities In The Industrial Internet Of Things, Mohammed Amin Almaiah, Fuad Ali El-Qirem, Rami Shehab, Khaled Sulieman Momani

Mesopotamian Journal of Computer Science

Data-driven decision-making, real-time connectivity, and automation have transformed industrial operations with the Industrial Internet of Things. However, the integration also introduces substantial cybersecurity vulnerabilities, making IIoT networks a prime target for malicious activities. Cyber threats are evolving and becoming more sophisticated, which makes traditional security mechanisms inadequate. An approach using deep learning to detect malicious activities in IIoT environments is examined. It is investigated whether Deep Feed Forward neural networks, autoencoders, and convolutional neural networks are effective at detecting anomalies and mitigating cyber threats. NSL-KDD and UNSW-NB15 benchmark datasets are used to evaluate the proposed model's accuracy, precision, and detection …

Self-Attention Enhanced Dual Bigru For Arabic Fake News Detection, Baqer M. Merzah, Jafar Razmara, Jaber Karimpour

Mesopotamian Journal of Computer Science

The rapid proliferation of social media platforms has greatly amplified the dissemination of fake news, representing significant obstacles to public trust and evidence-based decision-making, particularly for the Arabic-speaking population. Meeting the challenge of Arabic fake news detection is a problem compounded by the complex morphological nature of the language, as well as limited resources. This study presents a hybrid deep learning framework that integrates two Bidirectional Gated Recurrent Units (BiGRUs) along with an attention mechanism for efficiently detecting misinformation in Arabic news. The method leverages FastText word embeddings for disambiguating the intricate semantics of the Arabic language. The model is …

Haze-Image-Dataset: A Large-Scale Benchmark For Image Dehazing In Variable Fog And Low-Light Conditions, Mustafa J. Shahbaz, Ali A.D. Al-Zuky

Mesopotamian Journal of Computer Science

To enhance image dehazing and visual recognition in real-world conditions, we introduce HAZE-IMAGE-DATASET, a large-scale dataset comprising nearly 42,000 images. It is constructed from 1,532 clean images sourced globally and captured using a Samsung smartphone, covering diverse natural and urban scenes. The dataset includes synthetic and real haze variations. Synthetic haze was generated using MATLAB-based atmospheric scattering models with depth maps for 10 fog levels. Colored haze was created using alpha blending (α = 0.4) in six colors: red, green, blue, yellow, white, and black. Low-light conditions were simulated via uniform darkening at 10 levels. Also, 616 real haze images …

Deep Learning For Uav Classification: Impact Of Noise And Multipath Fading In Rf Signals, Prajoy Podder, Maciej Jan Zawodniok, Sanjay Kumar Madria

Electrical and Computer Engineering Faculty Research & Creative Works

The increasing presence of unmanned aerial vehicles (UAVs) raises serious security concerns, particularly regarding unauthorized drone operations. Recent U.S. security statistics report a sharp rise in unauthorized UAV activities, with the Federal Aviation Administration (FAA) receiving over 100 monthly reports of illegal drone operations near airports. In 2024 alone, Dedrone records 1.19 million unauthorized drone flights across major U.S. cities, highlighting the need for robust UAV detection and classification systems. In this work, a lightweight Convolutional Neural Network (CNN) model is proposed for RF-based UAV classification under noisy and multipath fading conditions. The proposed CNN consists of multiple convolutional blocks, …

Generalizing Medical Image Segmentation Task With Efficient Deep Learning Models, Abel A. Reyes-Angulo

Dissertations, Master's Theses and Master's Reports

Medical Image Segmentation is a critical task in the field of medical imaging, playing a crucial role in diagnostics, treatment planning, and disease monitoring. The emergence of Deep Learning (DL) has ushered in a new era in Artificial Intelligence (AI), propelling remarkable advancements in key domains like language translation, object recognition, and recommendation systems. This evolution has been accompanied by continuous enhancements in computational efficiency and improvements in predictive accuracy. The introduction of sophisticated algorithms, such as convolutional neural networks (CNNs) and transformers, exemplifies these advancements. DL algorithms have demonstrated exceptional efficacy in medical image segmentation tasks, showcasing the potential …

Applications Of Deep Learning For Optimizing Fingerphoto And Latent Fingerprint Biometrics, Amol Sanjay Joshi

Graduate Theses, Dissertations, and Problem Reports (ETD)

Fingerprint-based biometric recognition remains one of the most dependable and widely adopted approaches for identity verification due to its permanence and distinctiveness. Recent advancements in mobile and contactless imaging have extended fingerprint acquisition beyond controlled environments into unconstrained, real-world conditions through fingerphotos, contactless fingerprint images captured by digital or smartphone cameras. While this paradigm shift enhances accessibility and user convenience, it introduces significant technical challenges. Variations in illumination, focus, and motion blur often degrade ridge patterns, making accurate feature extraction and matching more difficult. Similarly, in forensic applications, latent fingerprints, incomplete or smudged prints lifted from surfaces, pose unique challenges …

Deep Learning And Adaptive Clustering Approaches For Flood Prediction And Efficient Sensor Placement In Missouri, Fahimeh Sharafkhani

Doctoral Dissertations

Floods represent formidable natural calamities, posing a significant threat to communities and infrastructure due to their unpredictable and often devastating consequences. The occurrence of floods is influenced by a convergence of meteorological, hydrological, and geographical factors, resulting in changes to the patterns of rising water levels. Machine learning models have emerged as favored tools in recent times for modeling water levels and enhancing the precision of flood predictions. This research employs both supervised and unsupervised machine learning models, with the main objective of improving the accuracy of flood predictions and sensor placement. Four distinct deep learning models are used to …

Leveraging Deep Learning Models And Social Media Data For Enhanced Situation Awareness In Disaster Management, Ademola Abdulganiyu Adesokan

Doctoral Dissertations

"In recent years, social media has become a crucial source of real-time data for disaster management, supporting emergency responses when traditional channels like 911 are overcrowded and overwhelmed. It offers authorities valuable data for developing effective strategies, especially when swift actions are essential to save lives. However, the informal language, ambiguous meanings, and irrelevant content on social media pose challenges to accurate classification and hinder the efficient extraction of disaster-relevant information, leading to inefficiencies in emergency response efforts.

This research focuses on seven key questions: i) How can we detect, classify, and analyze hate and offensive tweet emotions during large-scale …

Knowledge And Ontology Enhanced Approach To Natural Language Understanding (Koe-Nlu) In Computational Social Media And Healthcare, Naga Usha Gayathri Lokala

Theses and Dissertations

Natural 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 …

Leveraging Artificial Intelligent For Optimized Crop Production: An Ann-Based Approach, Yahya Layth Khaleel, Fadya A. Habeeb, Mustafa Abdulfattah Habeeb, Fatimah N. Ameen

Mesopotamian Journal of Computer Science

To incite modern day crop production and ensure sustainability, exact crop recommendations are key to the process. This study pays significant attention to the need for the use of big data tools in studies involving comprehensive data sets that contain information on soil and other environmental characteristics. The set of data used in this research includes Nitrogen, Phosphorus, and Potassium content coordinated with Temperature, Humidity, pH Value, and Rainfall. Knowing these factors is to make a favorable decision about improving agricultural products yield, availability and management of the resources, as well as general well-being of the crops. Specialized advisory on …

Enhancing Cybersecurity In Smart Education With Deep Learning And Computer Vision: A Survey, Guma Ali, Aziku Samuel, Maad M. Mijwil, Kholoud Al-Mahzoum, Malik Sallam, Ioannis Adamopoulos, Ayodeji Olalekan Salau, Indu Bala, Klodian Dhoska, Engin Melekoglu

Mesopotamian Journal of Computer Science

The rapid digital transformation of education, driven by the widespread adoption of smart devices and online platforms, has ushered in the era of smart education. While this shift enhances learning experiences, it also introduces significant cybersecurity risks that threaten the confidentiality, integrity, and availability of educational resources, student data, and institutional systems. This survey examines how deep learning (DL) and computer vision (CV) techniques can enhance cybersecurity in smart education environments. By reviewing 202 peer-reviewed research papers published between January 2022 and June 2025 across leading publishers such as ACM Digital Library, Frontiers, Wiley Online Library, IGI Global, Nature, Springer, …

Enhancing Image Classification Using A Convolutional Neural Network Model, Zena M. Saadi, Ahmed T. Sadiq, Omar Z. Akif, Marwa M. Eid

Journal of Soft Computing and Computer Applications

In recent years, with the rapid development of the current classification system in digital content identification, automatic classification of images has become the most challenging task in the field of computer vision. As can be seen, vision is quite challenging for a system to automatically understand and analyze images, as compared to the vision of humans. Some research papers have been done to address the issue in the low-level current classification system, but the output was restricted only to basic image features. However, similarly, the approaches fail to accurately classify images. For the results expected in this field, such as …

Real-Time Feedback-Driven Framework For Automated Cybersickness Mitigation, Md Jahirul Islam

Master's Theses

As technologies are becoming more advanced day by day, the embracement of virtual reality (VR) technology among users is also increasing in daily activities for various purposes, and subsequently, the barrier between the real and virtual world is fading. Despite the versatile uses, cybersickness (CS) is a major problem which is induced among users due to the immersive VR experience. There is a plethora of research findings and methods to measure the users’ CS such as virtual reality sickness questionnaire (VRSQ), simulator sickness questionnaire (SSQ), fast motion scale questionnaire (FMS), and others. Recently, machine learning approaches have also been adopted …

Identifying Subject Bias In Wifi-Based Human Activity Recognition Evaluation Methods, Amany Elkelany, Robert J. Ross, Susan Mckeever

Conference papers

WiFi-based Human Activity Recognition (HAR) has emerged as a promising approach for monitoring and analysing human activities in a non-intrusive manner, leveraging WiFi signals for activity classification. Despite advancements, existing WiFi-based HAR research lacks consideration of subject (human) bias. This results in learning models performing well on individuals used in the training samples but failing to generalise to new/unseen subjects, in contrast to known good practices in machine learning. In this paper, we address this oversight directly by systematically examining the evaluation methodology for the WiFi-based HAR context. Specifically, we investigate the impact of Leave-One-Subject-Out Cross-Validation (LOSOCV) in a hybrid …

Robust Learning With Probabilistic Relaxation Using Hypothesis-Test-Based Sampling, Zilin Wang

Dissertations and Theses Collection (Open Access)

In recent years, deep learning has been a vital tool in various tasks. The performance of a neural network is usually evaluated by empirical risk minimization. However, robustness issues have gained great concern which can be fatal in safety-critical applications. Adversarial training can mitigate the issue by minimizing the loss of worst-case perturbations of data. It is effective in improving the robustness of the model, but it is too conservative, and the plain performance of the model can be unsatisfying. Probabilistic Robust Learning (PRL) empirically balances the average- and worst-case performance while the robustness of the model is not provable …

Deep Learning Framework For Inverse Problems In Computational Imaging: A Lensless Imaging And Super-Resolution Magnetic Resonance Imaging Case, Arpan Poudel

Graduate Theses and Dissertations

Inverse problems in computer vision involve reconstructing an original scene or image from incomplete, noisy, or indirect measurements. These problems are critical in tasks such as image denoising, deblurring, super-resolution, and lensless imaging, where the goal is to recover high-quality images from degraded or partial measurements. This thesis introduces novel approaches to address two specific real-world inverse problems: (1) image super-resolution in medical imaging and (2) lensless image reconstruction . In the first part of this work, we tackle the problem of image super-resolution in Magnetic Resonance Imaging (MRI). High-resolution MRI scans are often limited by hardware constraints, patient movement, …

Cmos-Based Rotational Spectroscopy: Massive Spectral Fingerprint Generation And Molecular Detection With Deep Learning, Yasamin Fozouni

Computer Science and Engineering Theses and Dissertations

Rotational Spectroscopy is a powerful spectral fingerprinting approach that can be used for identifying different gas molecules in a sample. Gas molecules are free to rotate, with inertia, in fixed states of quantized energy. In Rotational Spectroscopy, radiative beams are shown onto a sample to cause an energy-based transition between quantized rotational states. By sweeping the frequency of the radiative beams and monitoring the absorption with a sensor, one can profile the different rotational states, monitoring for energy based transitions. These transitions are dependent on unique properties of the molecules, thus presenting a unique molecular identification fingerprint (in the form …

A Machine Learning Approach For Estimating Evapotranspiration For Urban Landscaping Vegetation In Semi-Arid Regions, Damian Lorenzo Gallegos Espinoza

Open Access Theses & Dissertations

Water management is important for residents in semi-arid urban areas due to increasing demand, water scarcity, and rising costs. It is estimated that in semi-arid regions, 40-70% of the household water consumption is used in landscaping. Therefore, urban landscaping water use can substantially contribute to water conservation. This work aims to estimate the water needs of urban landscaping vegetation to inform residents in semi-arid regions.Evapotranspiration indicates water and energy exchange between the atmosphere, soil, and vegetation. This interaction depends on solar radiation, evaporation, transpiration, and other biophysical parameters. Evapotranspiration has become a reference for water management in agriculture (e.g., crop …

Detecting Anomalies In Dynamic Attributed Graphs: An Unsupervised Learning Approach, Austin Hamilton

Electronic Theses and Dissertations

Dynamic attributed graphs, which evolve over time and hold node-specific attributes, are essential in fields like social network analysis, where anomalous node detection is a growing area. Vehicular social networks (VSNs), a subset of these graphs, are ad hoc networks in which vehicles exchange data with one another and with infrastructure. In this dynamic context, identifying anomalous nodes is challenging but crucial for maintaining trust within the network. This work presents an unsupervised deep learning approach for anomalous node detection in VSNs. This model achieved an accuracy of 71% while detecting synthetic anomalies in a simulated network based on real-world …

Real-Time Network Simulations For Ml/Dl Ddos Detection Using Docker, Luis D. Garcia

Master's Theses

As the integration of artificial intelligence (AI) within cybersecurity continues to

grow, machine learning (ML) and deep learning (DL) models are increasingly used to

detect cyber attacks. However, these models are rarely evaluated in real-time attack

scenarios to see how subtle changes from the real networking environment can affect

their predictions. To address this issue, we propose a scalable, platform-independent

Docker testbed specifically designed for simulating real-time Distributed Denial of

Service (DDoS) attack scenarios that allows researchers to deploy and evaluate their

pre-trained, ML and DL detection models. Our framework is simple to configure

and can run across Intel and …

Analyzing And Extending Machine Learning Frameworks On High Risk Domains, Chengbin Hu

USF Tampa Graduate Theses and Dissertations

Machine learning (ML) has become a transformative force in high-risk domains such as genomics and cybersecurity, where accurate predictions and robust defenses are essential. This dissertation advances ML frameworks in these areas by developing methods to enhance predictive power in health applications and assess vulnerabilities in machine learning systems.

In the genomics field, the work addresses challenges in Non-Invasive Prenatal Testing (NIPT) of monogenic disorders by proposing a deep learning model that reconstructs the fetal genome using maternal plasma cell-free DNA (cfDNA) and parental whole-genome sequencing (WGS) data. This model achieves high accuracy in single nucleotide variation (SNV) prediction, surpassing …

Neutrosophic Logic-Based Crop Yield Prediction And Risk Assessment Using Least Squares Regression, M. Srikanth, R.N.V. Jagan Mohan, M. Chandra Naik

Neutrosophic Systems with Applications

Agriculture faces significant challenges due to climate change and unpredictable environmental factors, which impact crop yields and threaten food security. This study proposes a novel approach to crop yield prediction and risk assessment using neutrosophic logic and least squares regression. By integrating these methods, we aim to improve accuracy in predicting crop losses under uncertain conditions. The model classifies crops based on profitability and environmental risks, utilizing the independence test to evaluate the relationships between crop attributes. Our approach leverages deep learning techniques, such as restricted Boltzmann machines (RBM), to enhance the analysis of crop data and provide farmers with …

Neutrosophic Logic-Based Crop Yield Prediction And Risk Assessment Using Least Squares Regression, M. Srikanth, R.N.V. Jagan Mohan, M. Chandra Naik

Neutrosophic Systems with Applications

Agriculture faces significant challenges due to climate change and unpredictable environmental factors, which impact crop yields and threaten food security. This study proposes a novel approach to crop yield prediction and risk assessment using neutrosophic logic and least squares regression. By integrating these methods, we aim to improve accuracy in predicting crop losses under uncertain conditions. The model classifies crops based on profitability and environmental risks, utilizing the independence test to evaluate the relationships between crop attributes. Our approach leverages deep learning techniques, such as restricted Boltzmann machines (RBM), to enhance the analysis of crop data and provide farmers with …