Digital Commons Network^™

Toward Reliable Computational Social Science: Inconsistency-Aware Methods For Human Annotation And Ai Inference, Sujan Dutta

Theses

As artificial intelligence (AI) becomes increasingly common in computational social science, \textit{inconsistency} has emerged as a key challenge. AI models often contradict themselves when given equivalent inputs, disagree with other models on the same data, and diverge from human judgments in seemingly opaque ways. Human annotators exhibit their own inconsistencies, both within individuals and across groups shaped by differing values and identities. Rather than treating these inconsistencies simply as noise, this dissertation argues that they contain meaningful signals that can be leveraged to improve learning efficiency, strengthen evaluation, and increase the reliability of large-scale social measurement. To study this phenomenon, …

What Do Future Nutrition Professionals Think About Wic Jobs? A Study Of Nutrition Career Perceptions, Madison Degenfelder

Theses

Background: The Special Supplemental Nutrition Program for Women, Infants and Children (WIC) program faces a critical workforce shortage, yet little is known about the knowledge, attitudes, and beliefs of nutrition students and recent graduates regarding WIC career pathways, requirements, and compensation. Understanding these factors is essential to design strategies to strengthen recruitment and build a sustainable WIC workforce. Objective: To identify and describe knowledge, attitudes and beliefs on entering the WIC workforce among college nutrition and dietetics students and recent graduates. Design: A mixed-methods, sequential study design which included a cross-sectional survey and two focus groups was employed to ascertain …

'Guide The Tide' - Supervision Of International Medical Graduates In The Emergency Department: A Mixed Methods Study, Purnasankar Bhowmik

Theses

Australia’s emergency departments (EDs) have experienced significant stress since the COVID-19 pandemic. An increase in the number of patient presentations, a lack of in-patient beds for the disposition of admitted patients and long waiting times are contributing factors jeopardising patient safety. In addition, the shortage of skilled staff to look after patients in the ED is the paramount concern currently and moving forward. Similarly to many other specialty areas of medicine, EDs rely on international medical graduates (IMGs) to fill the gaps in the medical workforce. However, concerns have been raised regarding IMG diversity in training requirements, clinical skills and …

Enhancing Json Schema Inference Via Semantic And Structural Modeling Of Heterogeneous Data, Justin R. Namba

Theses

Schema discovery is finding the structure of data. It helps users understand the meaning of data and write queries to manipulate it. This is typically easy for relational databases, but complex for non-relational (NoSQL) databases with documents. For relational databases, the schema is predefined because the data they contain is structured, but for NoSQL databases, data is usually unstructured or semi-structured. Here, we focus on a type of semi-structured data called JSON, which is a collection of documents that consists of nested key-value pairs. A JSON key and value are similar to a column name and its associated data instance …

Drift Aware Continual Finetuning Of Encoder-Only Text Classifiers Under Temporal Distribution Shift, Simran Kc

Theses

The performance of encoder-only text classifiers degrades silently as their input distribution drifts over time. This thesis presents a closed-loop pipeline for drift-aware continual fine-tuning that combines a sample size invariant drift severity metric, an automated retraining loop on the most drifted samples, and a validation gate that rejects unsafe updates before deployment. We evaluate the pipeline on two datasets, Amazon Electronics Reviews and HuffPost News Category, and observe significant performance improvements over the initially trained model, along with a substantial reduction in compute and labelling costs compared to fresh-per-year and cumulative retraining baselines.

Uae Traditional Folklore Through Modern Media And Its Educational Potential: The Case Of Freej, Shahad Imad Mostafa

Theses

This mixed‑methods study examined how the Emirati animated series Freej preserves traditional folklore through vernacular language, dialect, and idiomatic expressions, and evaluated its educational potential. Guided by four questions on folklore representation, dialect/idiom use, the role of the characters, and curricular applicability, the research combined content analysis of selected episodes, interviews coding and a survey of 100 Emirati participants; survey data were summarized with descriptive statistics to triangulate qualitative insights.

Findings indicated that audiences overwhelmingly perceived Freej as linguistically and culturally authentic: 95% rated the dialect as very/somewhat authentic, 94% said it reflects elders’ speech, and 89% felt the series …

Evaluating Datasets And Models For Offensive Language Identification, Skye Deson Morgan

Theses

Social media platforms enable people to easily communicate with one another, share ideas and express their opinions. Unfortunately, the anonymity provided by platforms such as Twitter and Facebook is more than enough to foster the spread of harmful content. Government agencies, as well as technology companies, are actively working towards finding better means of curbing offensive and objectionable content in social media and contributing to a more welcoming online environment. In this study, I first present a critical literature review on offensive language identification, exploring offensive language datasets and computational models using both machine learning and deep learning. Next, I …

Scout 07, Austin Roberts

Theses

Throughout my tenure at RIT, I have produced multiple short films, undertaken gaming co-ops, and pursued an independent study centered on gaming assets to expand my technical and creative skills. For my thesis film, I sought to merge these experiences by concentrating on environmental storytelling—a deliberate challenge, given faculty concerns that narratives lacking strong character focus may struggle to engage audiences emotionally. With these considerations in mind, I crafted a film featuring a secondary character within a rich environment, designed to immerse viewers and provoke curiosity about the story’s temporal and spatial context.

The Impact Of Screen Time Patterns On Digital Behaviour And Productivity: A Data Analytics Study Across Age Groups, Maytha Alshamsi

Theses

The increasing incidence of screen time of all ages has generated a greater interest in the effect of screens on the human psyche. This paper looks at behavioural data of 2, 000 individuals aged 13-64 to investigate the predictability of a composite mental- health score by various modalities of screen-time, lifestyle and psychological charac- teristics. We determine, based on exploratory statistics, multivariate regression and su- pervised machine-learned models, that in as much as the influence of total screen time and social-media consumption has significant relationships with worse mental health, the effects are small when juxtaposed to those of lifestyle and …

Comparative Analysis Of Machine Learning Models For Spam Email Detection, Rashed Almarri

Theses

The current research examines one of the most effective approaches to spam email detection based on machine learning and natural language processing (NLP). The study is placed in the context of the rising cyber threats and the influx of emails, where the spam/ham data is to be classified correctfully using the combination of the Logistic Regression, NLP (including tokenization, lemmatization, and TF-IDF vectorization). The questions of the research were devoted to the efficiency of such an approach and the interpretation of its results. The data used are obtained by a publicly available Kaggle data set that contains 5,572 labeled email …

Security Vulnerabilities In Cloud Storage: A Comparative Study Of Google Drive, Dropbox, And Onedrive, Sultan Majid Alshamsi

Theses

In recent years, use of internet has increased manifold. People have started creating more data, and instead of saving it locally, people have started preferring storing it “online”. Companies are also following the same path to maintain accessibility and availability of their data. Thus, data has become a central part of our lives. People, companies, and institutions rely heavily on cloud storage solutions these days for managing their data. Some of the main cloud storage solution are Google Drive, Dropbox, and OneDrive. These are sophisticated solutions developed by tech giants, and general expectation of the public is that these solutions …

Exploring The Relationship Between Social Media And Anxiety In Deaf Community Students: A Mixed-Methods Approach., Gigi Zheng

Theses

The mental health of Deaf individuals, particularly in relation to social media engagement, is an understudied area amid growing global concerns about anxiety disorders. Deaf individuals face unique challenges, including communication barriers, social isolation, and stigma, all of which can heighten anxiety levels. Although social media offers valuable opportunities for connection and self-expression, it may also amplify stress and feelings of inadequacy when accessibility barriers persist. This mixed-methods study examined anxiety among Deaf college students and explored the cultural validity of two widely used anxiety measures—the State-Trait Anxiety Inventory (STAI) and the Beck Anxiety Inventory (BAI). Quantitative data were collected …

Beyond The Mozart Effect: How Musical Experience And Emotional Response Shape Individual Differences In Spatial Reasoning, Connor Watkinson

Theses

Research on the Mozart effect (the short-term enhancement of spatial-temporal reasoning after music listening) has yielded inconsistent results, often due to overlooking individual differences and using limited stimuli. This study examined multiple musical conditions (popular lyrical, classical, and white noise control) while considering the moderating roles of musical experience and the mediating influence of emotional responses. Undergraduate participants (N = 162) completed spatial reasoning assessments before and after an eight-minute listening session, alongside measures of musical background, listening habits, mood, and music-evoked emotions.Analyses revealed no overall difference in spatial reasoning improvement between music and white noise conditions. However, substantial individual …

Detecting Fraud In Police Reports Using Machine Learning And Natural Language Processing, Maryam Almarar

Theses

The paper explores how statistical analysis and machine learning can be used to identify the fraud patterns in the police reports. The study aims at establishing the most important predictive factors and indicators distinguishing fraudulent and valid cases with the use of structured data of police databases. The work was done in the background of the increase in financial fraud instances and the rising necessity of the introduction of automated detection systems in police departments. Police reports of the pastwere mined down to data and analyzed on SPSS 1, to carry out statistical operations. The sample was structured data which …

Segment-Level Machine Learning For Detecting Partial Deepfake Audio: An Rnn–Svm Hybrid Approach For Real-World Adversarial Environments, Ahmad Alhelli

Theses

This study investigates the detection of real, fully fake, and partially fake (PF) audio using classical machine-learning models as well as a segment-level analysis model. Unlike most existing research, which focuses solely on binary real-vs-fake classification, this work introduces a three-class detection framework and constructs realistic PF samples by inserting short synthetic speech segments into otherwise genuine audio recordings. Its method combines MFCC and spectral feature engineering,Wav2Vec2 embeddings, controlled PF synthesis and various models such as SVM, Random Forest, XGBoost and an attention based RNN. Segment level windowing allows fine-grained study of transition of time and breaks of manipulation. The …

Rise Of Social Media Hacking: Ai-Based Ip Tracking For Uae Law Enforcement, Mounikha Naarrayen Chakravarthula

Theses

Social media has evolved into a critical channel for communication, expression, and public influence, but it has also become a prevalent avenue for cybercrime, particularly in digitally advanced nations such as the United Arab Emirates (UAE). The rising complexity of online offences, coupled with anonymisation tools and cross-border digital behaviour, has made the attribution of social-media-based cyber incidents increasingly challenging for law enforcement. In this context, artificial intelligence (AI) offers the potential to strengthen digital investigations by providing intelligent, scalable, and evidence-driven attribution capabilities. This research develops an AI-assisted Internet Protocol (IP) attribution framework tailored specifically for UAE law enforcement …

Predictive Ai Models For Phishing Attack Detection: A Data-Driven And Statistical Analysis Approach, Suhail Othman Alfalasi

Theses

Phishing has remained a serious threat to cybersecurity, as this type of attack can easily bypass detection systems that are either rule-based or blacklist-based. The proposed thesis work will present a solution that will make use of both statistical analysis and machine learning to correctly identify a phishing website. The solution will make use of a hybrid approach comprising statistical-based preprocessing methodologies, such as PCA and decision tree-based feature selection, to filter the crucial URL features that a website may possess. A carefully balanced dataset has been utilized, as well as a non-parametric approach utilizing the Mann-Whitney U-test to validate …

Multilingual Identity Document Information Extraction Via Dynamic Templates And Hybrid Ocr, Khalifa Rashed

Theses

The thesis introduces a unified, automated document intelligence system that can in-house resolve significant issues in identity authentication and data OCR by two complementary technological components: advanced facial biometrics and powerful multilingual optical character recognition. The system directly addresses the inefficiencies of the slow speed of document processing by humans, human error and high cost of operation by offering a single pipeline for verifying the identity of the user via facial comparison and digitizing textual data contained in documents. The former module adopts an advanced face verification pipeline. It starts with an image pre-processing step, which improves the quality of …

Assessing The Impact Of Codec-Induced Audio Degradation On Voice Biometric Systems, Suhil Ali Almuhaisni

Theses

This study examines the robustness of voice biometrics when speech signals undergo audio codec transformations and sampling rate variations, conditions common in telecommunication networks. Speaker verification systems such as ECAPA-TDNN perform well on clean datasets, but their accuracy declines when low-bitrate codecs compress speech or when signals are resampled at reduced frequencies. In real-world deployments, systems adapt audio to bandwidth and storage limitations, often removing subtle acoustic details that support consistent speaker recognition. The research will analyse how codec settings and sampling rates, particularly those optimized for efficiency in bandwidth-limited systems, influence the stability of speaker embeddings. Instead of ranking …

Beyond Detection: A Batch-Based Ai Framework For Temporal And Event-Correlated Trend Analysis Of Misinformation On Social Media, Vishnu Tejas Vijayaraghavan

Theses

The rapid spread of misleading information on social media influences public behaviour and complicates crisis communication. Although transformer models such as BERT accurately detect misinformation at the post level, most studies analyse posts in isolation and overlook howmisinformation fluctuates over time or responds to major events. This study addresses that gap by developing an end-to-end analytical workflow that integrates BERT-based classification with temporal aggregation, topic clustering, anomaly detection, and event alignment. The analysis uses 10,700 COVID-19–related tweets (6,420 training, 2,140 validation, and 2,140 testing). Because timestamps were unavailable, synthetic timestamps were assigned using an evenly spaced date range between 1 …

Assessing Teachers’ Practices In Supporting Gifted/Talented And Twice-Exceptional Students In Schools And Centers In Abu Dhabi, Zayed Jaber

Theses

Gifted/talented and twice-exceptional (2e) students, especially those whose needs are often overlooked in mainstream classrooms, are more likely to benefit from consistently implemented Differentiated Instructional Practices (DIPs). This study aimed to examine implelemntation of DIPs for gifted/talented and 2e students and to compare the practices implemented between of Special Educational Needs (SEN) teachers and General Education teachers in schools and centers within Abu Dhabi. This research used a quantitative, cross-sectional survey design to collect data from eighty-six teachers from Abu Dhabi. A questionnaire was designed to rate teachers’ self-reported implementation of DIPs across the four domains of content, process, product, …

Enhancing The Acceptability Of Decision-Making System Using Xai Case: Crime Profiling System, Mohamed Saeed Al Shamsi

Theses

In the current world, we need to place more emphasis on how easily interpretable, accurate, and acceptable data analysis results are, given that essential operations in law enforcement, among other sectors, are backed up by the use of complex computing systems. Crime profiling systems that use crime data for profiling encounter major problems because they depend on algorithm-based methods. These methods can be ambiguous and inaccurate, leading to low public acceptability. The study investigates major problems with Complex Crime profiling systems (CPS) because their unexplained algorithms result in system performance issues and public scepticism. XAI provides a solution to handle …

It Hurts To Become, Ana Maria Joyce

Theses

Using the medium of oil painting, this thesis will investigate the ways that memories of my past inform and distort my perception of reality. Using domestic spaces as a framework in combination with my own body as the subject matter, I am creating compositions that redefine the meaning of home as a concept. Rather than a physical place of comfort, home is a cemetery for past selves that I revisit in my mind and grieve, no matter the time that has passed. By expressing this internal decay through physically scraping and sludging paint onto large scale surfaces, I aim to …

Ambiguous Bodies: Third Gender Expressions In Ancient Maya Art, Jayme Horne

Theses

This thesis examines the interpretations of Lintels 23, 24, and 25 from Yachilán, Drawing 18 from Naj Tunich, and Stela H from Copán, with a specific focus on third gender expression. These artworks depict figures that blend masculine and feminine attributes, from costuming to actions, to present intentionally ambiguous representations of the body that all ancient Maya people would have understood. Third gender expression is defined through the blending of masculine and feminine signifiers. It is a gender identity for people who do not identify as male or female, but rather as neither, both, or a combination of male and …

Enhancing Llm Code Generation: A Systematic Evaluation Of Multi-Agent Collaboration And Runtime Debugging For Improving Accuracy, Reliability, And Latency, Nazmus Ashrafi

Theses

The use of large language models (LLMs) for automated code generation has emerged as a significant focus within AI research. As these pretrained models continue to evolve, their ability to understand and generate complex code structures has opened up new possibilities for automating intricate programming tasks with greater accuracy. Although contemporary foundational models demonstrate promising results, researchers continue to explore optimal post-training strategies to enhance code quality. These include supervised fine-tuning, retrieval-augmented generation (RAG), debugging, and many others. In this thesis, I combine two such widely used post training approaches—namely (1) multi agent collaboration and (2) runtime execution of information-based …

Making Math Accessible: Addressing Language Challenges In 3rd Grade English Medium Math Classrooms, Salha Mohamad

Theses

The research article, Making Math Accessible: Addressing Language Difficulties in 3^rd Grade English Medium Math Classrooms, addresses the issue of how the teachers in UAE elementary schools assist a student who is to study mathematics in English rather than Arabic. In the School X, Grades 1 and 2 applied mathematics and science using the Ministry of Education (MOE) Arabic-medium curriculum. Since Grade 3, however, the school has used the American curriculum, so students are expected to study these subjects in English. This change brings major barriers in understanding basic mathematical terms, word problems, and instructions. The research applies …

Improving The Accessibility Of Speech For Deaf And Hard-Of-Hearing Individuals Through Affective Captions, Caluã De Lacerda Pataca

Theses

Captions have traditionally served as a bridge between the spoken word and its written representation, helping make speech accessible to Deaf and Hard-of-Hearing (DHH) individuals. It is worth considering, however, how much from speech is left out by this ’bridging’ between sound and visuals. This dissertation describes a research project that has, over six studies, looked at this very issue. We first examined whether there is an issue here at all. What does the experience of DHH individuals with captioning systems tell us about these systems’ shortcomings? For one, we found, captions are felt as monotonous and ambiguous. While communication …

Democratizing Community Discourse Analysis In Computational Social Science, Md Towhidul Absar Chowdhury

Theses

Community resource inequities and undetected infrastructure vulnerabilities cost municipalities billions annually, with disproportionate impacts on marginalized communities. Current computational approaches to community needs assessment suffer from two critical limitations: they rely on aggregate-level analysis that obscures granular community expressions, and they implement sophisticated computational tools that remain inaccessible to stakeholders without technical expertise. This creates what we term the Community-Computational Gap—a persistent divide where domain experts with vital contextual knowledge cannot access the analytical tools they need, while computational experts develop models without sufficient community context. This dissertation addresses these challenges through a novel methodological framework for fine-grained utterance-level classification …

Dyadic And Multiparty Turn-Taking In Two Deaf Preschool Classrooms, Savannah Tellander

Theses

Conversational turn-taking skills are needed to succeed in most conversational interactions. Timing of turn-taking is crucial for young deaf signers, who need early, accessible, and consistent input in a signed and/or spoken language. For deaf children who acquire a language in either modality, they must acquire appropriate turn-taking skills. Without these abilities, children may miss critical language input, and in a cascading effect–they may miss opportunities both for input (and thus new knowledge) and expressive language. In addition to documenting deaf children’s turn-taking skills across development, it is important to also document timing data for further understanding of cognitive language …

Digital Biophilia In Web Interface Design For Accessibility And User Well-Being, Bryana Michelle Peifer

Theses

This interdisciplinary study explores the intersection between horticultural therapy and web design, focusing on the impact of biophilic design principles in shaping contemporary practices within digital interface design. This exploration addresses the multifaceted relationship between humanity and the natural world, particularly how this connection serves as a source for the emergence of nature-inspired elements in digital expressions and their subsequent integration into modern web design for accessibility and well-being. As digital environments increasingly dominate daily experiences, the need for digital wellness through nature integration has become critical, especially for individuals who have physical disabilities, limited mobility, or limited access to …