Computational Linguistics Commons^™

The Security Of Llm-Generated Code, Christopher Brian Gonzalez Ayala

Student Theses

The rapid adoption of Large Language Models (LLMs) in software development has transformed coding practices by enabling automated code generation, completion, and optimization. Despite these advantages, concerns persist regarding the security and reliability of LLM-generated code. This study presents a comprehensive evaluation of both the functional correctness and security of code produced by three prominent LLMs as of early 2026. A total of 4,800 code snippets were generated using 100 security-focused programming prompts derived from the OWASP Top 10:2025, translated across eight natural languages and two phrasing styles (literal and natural developer-oriented prompts). To assess performance, a multi-stage experimental framework …

Interpreting American Sign Language: A Literature Review Of Assistive Technologies, Natalie Louise Paradiso, Emma Grace Kochenderfer

Student Scholar Symposium Abstracts and Posters

American Sign Language (ASL) is a visually elaborate, spatially oriented linguistic methodology that relies on combinations of hand movements, body positioning, facial expressions, and motion/spatial perception, aspects of which make interpretation difficult for automated machine recognition. Current assistive technology approaches to ASL interpretation are generally within the categories of computer vision models (including deep learning, multi-focus image fusion, and keypoint tracking) and wearable, multimodal/sensor-based approaches (such as smart glasses and inertial-sensor gloves). Within controlled environments, computer vision models perform well. However, when applied to conditions such as non-manual signs/features, signer variability, and rapid assimilation, they falter in processing all aspects …

Propasafe-Hybrid: A Text-Based Hybrid Propaganda Detection Tool, Thomas Kimmeth, Avijit Roy, Vivek Sharma

Publications and Research

Propagandistic content increasingly circulates through online news and social media, where readers often encounter it with limited scrutiny, highlighting the need for reliable and fine-grained detection. This paper introduces Propasafe-Hybrid, a sentence-level system that integrates a fine-tuned transformer classifier with LLM-based technique classification to identify, label, and explain specific propaganda strategies. The pipeline generates actionable outputs, including highlighted sentences, technique assignments, and concise rationales, so users can immediately understand why a sentence was flagged and how each label was determined. To control inference cost, Propasafe-Hybrid employs a cost-aware pre-filtering stage that forwards only high-likelihood sentences to LLMs, reducing token usage …

Propasafe-Hybrid: A Text-Based Hybrid Propaganda Detection Tool (Ila 2026 Presentation), Thomas Kimmeth, Avijit Roy, Vivek Sharma

Publications and Research

This presentation introduces Propasafe-Hybrid, a hybrid system for sentence-level propaganda detection that combines offline transformer-based classification with selective large language model (LLM) explainability. The system employs a two-stage pipeline in which a local BERT-based classifier evaluates all input text and filters non-propagandistic content, while only high-confidence candidates are forwarded to an LLM for rhetorical technique labeling and explanation. This design enables cost-aware, privacy-conscious, and scalable analysis by reducing unnecessary reliance on external models.

Propasafe-Hybrid identifies propagandistic techniques such as loaded language, obfuscation, and appeal to fear, and generates concise natural language rationales that make these techniques interpretable to users. By …

Structural Silence: When Ai Infrastructure Fails Speakers Of Underrepresented Languages, Avijit Roy, Proma Roy

Publications and Research

Artificial intelligence tools for education and language support are increasingly framed as scalable responses to access gaps in under-resourced communities. Yet the infrastructure underlying these tools—training corpora, tokenization schemes, evaluation benchmarks, and deployment architectures—encodes a set of assumptions that systematically disadvantages speakers of underrepresented languages before a single model is trained. This paper examines those assumptions through the lens of Bengali, one of the world’s most widely spoken languages with roughly 285 million speakers (Ethnologue, 2025; International Communication and Leadership School, 2026), and the structural barriers that emerge when attempting to build AI-assisted educational tools for Bengali-speaking learners in low-connectivity …

Syntax-Enhanced Boundary-Aware Named Entity Recognition Model, Chuanming Yu, Bin Deng, Zhengang Zhang

Journal of Scientific Information Research

[Purpose/significance] This study addresses the issue of inadequate perception of entity boundaries in traditional character-level modeling-based named entity recognition models by integrating syntax information containing entity boundary features into the task using a multi-head graph attention network with dense connections. This integration enhances the effectiveness of named entity recognition.

[Method/process] This study proposes a Syntax-enhanced Boundary-aware Named Entity Recognition Model (SynBNER), which utilizes BERT for text semantic representation and integrates syntax information using a dense-connected graph attention network. This integration incorporates implicit entity boundary information from syntax information into word representations, thereby enhancing the model's entity boundary perception capability.

[Result/conclusion] …

Revitalization Of Endangered Languages With Ai, Ivory Yang

Dartmouth College Master’s Theses

The preservation and revitalization of endangered languages, particularly those with minimal digital presence, presents significant challenges for computational linguistics. This thesis addresses these challenges by proposing novel methods for language identification and data generation, focusing on underrepresented Indigenous languages, specifically Nüshu, Native American and Native Alaskan languages.

In the first study, a COLING 2025 paper, we present NüshuRescue, an AI-driven framework designed to facilitate the preservation of Nüshu, an endangered script used exclusively by Yao women in China. Using minimal seed data, we demonstrate how GPT-4-Turbo can generate new translations, expanding a publicly available Nüshu-Chinese corpus, achieving 48.69% accuracy in …

A Proposed Ehrenfeucht-Fraïssé Game Model For Natural Language Processing Generative Adversarial Networks, Don Li

Anthós

Large Language Models (LLM’s) (e.g., ChatGPT) constitute both a significant research area and commercial application of AI. Current major LLM’s are built on Generative Pre-Trained Transformer (GPT) neural network architecture to perform natural language processing (NLP) tasks. Generative Adversarial Network (GAN) is another popular neural network architecture, which leverages a zero-sum game between constituent neural networks within the architecture to train the GAN, and is widely used for visual data applications. This article proposes a new GAN architecture for NLP: an EF-GAN whose underlying algorithm uses Ehrenfeucht–Fraïssé (EF) games, a game-theoretic approach from model theory to determine elementary equivalence of …

Using Ai For Qualitative Labeling: Consistency And Comparisons, James Temple

Honors Program Theses

This paper continues research that evaluates the capacity of artificial intelligence (AI) to perform qualitative coding tasks. The previous study found that AI models lacked consistency with themselves and did not agree with human coded data. Since that study, AI’s general level of intelligence has increased. Hence, this study re-evaluates how well the newest set of AI models (Claude 3 and Gemini) can perform qualitative coding tasks. When tested, the new AI models perform about the same or better than previous models depending on the metric tested. While Gemini and Claude 3 do not agree with human output any more …

Guilty Machines: On Ab-Sens In The Age Of Ai, Dylan Lackey, Katherine Weinschenk

Critical Humanities

For Lacan, guilt arises in the sublimation of ab-sens (non-sense) into the symbolic comprehension of sen-absexe (sense without sex, sense in the deficiency of sexual relation), or in the maturation of language to sensibility through the effacement of sex. Though, as Slavoj Žižek himself points out in a recent article regarding ChatGPT, the split subject always misapprehends the true reason for guilt’s manifestation, such guilt at best provides a sort of evidence for the inclusion of the subject in the order of language, acting as a necessary, even enjoyable mark of the subject’s coherence (or, more importantly, the subject’s separation …

Executive Order On The Safe, Secure, And Trustworthy Development And Use Of Artificial Intelligence, Joseph R. Biden

Copyright, Fair Use, Scholarly Communication, etc.

Section 1. Purpose. Artificial intelligence (AI) holds extraordinary potential for both promise and peril. Responsible AI use has the potential to help solve urgent challenges while making our world more prosperous, productive, innovative, and secure. At the same time, irresponsible use could exacerbate societal harms such as fraud, discrimination, bias, and disinformation; displace and disempower workers; stifle competition; and pose risks to national security. Harnessing AI for good and realizing its myriad benefits requires mitigating its substantial risks. This endeavor demands a society-wide effort that includes government, the private sector, academia, and civil society.

My Administration places the highest urgency …

Evaluating Neural Networks As Cognitive Models For Learning Quasi-Regularities In Language, Xiaomeng Ma

Dissertations, Theses, and Capstone Projects

Many aspects of language can be categorized as quasi-regular: the relationship between the inputs and outputs is systematic but allows many exceptions. Common domains that contain quasi-regularity include morphological inflection and grapheme-phoneme mapping. How humans process quasi-regularity has been debated for decades. This thesis implemented modern neural network models, transformer models, on two tasks: English past tense inflection and Chinese character naming, to investigate how transformer models perform quasi-regularity tasks. This thesis focuses on investigating to what extent the models' performances can represent human behavior. The results show that the transformers' performance is very similar to human behavior in many …

Predicting High-Cap Tech Stock Polarity: A Combined Approach Using Support Vector Machines And Bidirectional Encoders From Transformers, Ian L. Grisham

Electronic Theses and Dissertations

The abundance, accessibility, and scale of data have engendered an era where machine learning can quickly and accurately solve complex problems, identify complicated patterns, and uncover intricate trends. One research area where many have applied these techniques is the stock market. Yet, financial domains are influenced by many factors and are notoriously difficult to predict due to their volatile and multivariate behavior. However, the literature indicates that public sentiment data may exhibit significant predictive qualities and improve a model’s ability to predict intricate trends. In this study, momentum SVM classification accuracy was compared between datasets that did and did not …

Chatgpt As Metamorphosis Designer For The Future Of Artificial Intelligence (Ai): A Conceptual Investigation, Amarjit Kumar Singh (Library Assistant), Dr. Pankaj Mathur (Deputy Librarian)

Library Philosophy and Practice (e-journal)

Abstract

Purpose: The purpose of this research paper is to explore ChatGPT’s potential as an innovative designer tool for the future development of artificial intelligence. Specifically, this conceptual investigation aims to analyze ChatGPT’s capabilities as a tool for designing and developing near about human intelligent systems for futuristic used and developed in the field of Artificial Intelligence (AI). Also with the helps of this paper, researchers are analyzed the strengths and weaknesses of ChatGPT as a tool, and identify possible areas for improvement in its development and implementation. This investigation focused on the various features and functions of ChatGPT that …

Creating Data From Unstructured Text With Context Rule Assisted Machine Learning (Craml), Stephen Meisenbacher, Peter Norlander

School of Business: Faculty Publications and Other Works

Popular approaches to building data from unstructured text come with limitations, such as scalability, interpretability, replicability, and real-world applicability. These can be overcome with Context Rule Assisted Machine Learning (CRAML), a method and no-code suite of software tools that builds structured, labeled datasets which are accurate and reproducible. CRAML enables domain experts to access uncommon constructs within a document corpus in a low-resource, transparent, and flexible manner. CRAML produces document-level datasets for quantitative research and makes qualitative classification schemes scalable over large volumes of text. We demonstrate that the method is useful for bibliographic analysis, transparent analysis of proprietary data, …

Toward Suicidal Ideation Detection With Lexical Network Features And Machine Learning, Ulya Bayram, William Lee, Daniel Santel, Ali Minai, Peggy Clark, Tracy Glauser, John Pestian

Northeast Journal of Complex Systems (NEJCS)

In this study, we introduce a new network feature for detecting suicidal ideation from clinical texts and conduct various additional experiments to enrich the state of knowledge. We evaluate statistical features with and without stopwords, use lexical networks for feature extraction and classification, and compare the results with standard machine learning methods using a logistic classifier, a neural network, and a deep learning method. We utilize three text collections. The first two contain transcriptions of interviews conducted by experts with suicidal (n=161 patients that experienced severe ideation) and control subjects (n=153). The third collection consists of interviews conducted by experts …

Exploring The Personality Of Virtual Tutors In Conversational Foreign Language Practice, Johanna Dobbriner, Cathy Ennis, Robert J. Ross

Conference papers

Fluid interaction between virtual agents and humans requires the understanding of many issues of conversational pragmatics. One such issue is the interaction between communication strategy and personality. As a step towards developing models of personality driven pragmatics policies, in this paper, we present our initial experiment to explore differences in user interaction with two contrasting avatar personalities. Each user saw a single personality in a video-call setting and gave feedback on the interaction. Our expectations, that a more extroverted outgoing positive personality would be a more successful tutor, were only partially confirmed. While this personality did induce longer conversations in …

Do It Like A Syntactician: Using Binary Gramaticality Judgements To Train Sentence Encoders And Assess Their Sensitivity To Syntactic Structure, Pablo Gonzalez Martinez

Dissertations, Theses, and Capstone Projects

The binary nature of grammaticality judgments and their use to access the structure of syntax are a staple of modern linguistics. However, computational models of natural language rarely make use of grammaticality in their training or application. Furthermore, developments in modern neural NLP have produced a myriad of methods that push the baselines in many complex tasks, but those methods are typically not evaluated from a linguistic perspective. In this dissertation I use grammaticality judgements with artificially generated ungrammatical sentences to assess the performance of several neural encoders and propose them as a suitable training target to make models learn …

Synthetic, Yet Natural: Properties Of Wordnet Random Walk Corpora And The Impact Of Rare Words On Embedding Performance, Filip Klubicka, Alfredo Maldonado, Abhijit Mahalunkar, John D. Kelleher

Conference papers

Creating word embeddings that reflect semantic relationships encoded in lexical knowledge resources is an open challenge. One approach is to use a random walk over a knowledge graph to generate a pseudo-corpus and use this corpus to train embeddings. However, the effect of the shape of the knowledge graph on the generated pseudo-corpora, and on the resulting word embeddings, has not been studied. To explore this, we use English WordNet, constrained to the taxonomic (tree-like) portion of the graph, as a case study. We investigate the properties of the generated pseudo-corpora, and their impact on the resulting embeddings. We find …

Perception & Perspective: An Analysis Of Discourse And Situational Factors In Reference Frame Selection, Robert J. Ross, Kavita E. Thomas

Conference papers

To integrate perception into dialogue, it is necessary to bind spatial language descriptions to reference frame use. To this end, we present an analysis of discourse and situational factors that may influence reference frame choice in dialogues. We show that factors including spatial orientation, task, self and other alignment, and dyad have an influence on reference frame use. We further show that a computational model to estimate reference frame based on these features provides results greater than both random and greedy reference frame selection strategies.

Multimodal Depression Detection: An Investigation Of Features And Fusion Techniques For Automated Systems, Michelle Renee Morales

Dissertations, Theses, and Capstone Projects

Depression is a serious illness that affects a large portion of the world’s population. Given the large effect it has on society, it is evident that depression is a serious health issue. This thesis evaluates, at length, how technology may aid in assessing depression. We present an in-depth investigation of features and fusion techniques for depression detection systems. We also present OpenMM: a novel tool for multimodal feature extraction. Lastly, we present novel techniques for multimodal fusion. The contributions of this work add considerably to our knowledge of depression detection systems and have the potential to improve future systems by …

Does The Test Work? Evaluating A Web-Based Language Placement Test, Avizia Long, Sun-Young Shin, Kimberly Geeslin, Erik Willis

Faculty Publications

In response to the need for examples of test validation from which everyday language programs can benefit, this paper reports on a study that used Bachman’s (2005) assessment use argument (AUA) framework to examine evidence to support claims made about the intended interpretations and uses of scores based on a new web-based Spanish language placement test. The test, which consisted of 100 items distributed across five item types (sound discrimination, grammar, listening comprehension, reading comprehension, and vocabulary), was tested with 2,201 incoming first-year and transfer students at a large, Midwestern public university. Analyses of internal consistency and validity revealed the …

Exploring The Functional And Geometric Bias Of Spatial Relations Using Neural Language Models, Simon Dobnik, Mehdi Ghanimifard, John D. Kelleher

Conference papers

The challenge for computational models of spatial descriptions for situated dialogue systems is the integration of information from different modalities. The semantics of spatial descriptions are grounded in at least two sources of information: (i) a geometric representation of space and (ii) the functional interaction of related objects that. We train several neural language models on descriptions of scenes from a dataset of image captions and examine whether the functional or geometric bias of spatial descriptions reported in the literature is reflected in the estimated perplexity of these models. The results of these experiments have implications for the creation of …

Data-Driven Synthesis And Evaluation Of Syntactic Facial Expressions In American Sign Language Animation, Hernisa Kacorri

Dissertations, Theses, and Capstone Projects

Technology to automatically synthesize linguistically accurate and natural-looking animations of American Sign Language (ASL) would make it easier to add ASL content to websites and media, thereby increasing information accessibility for many people who are deaf and have low English literacy skills. State-of-art sign language animation tools focus mostly on accuracy of manual signs rather than on the facial expressions. We are investigating the synthesis of syntactic ASL facial expressions, which are grammatically required and essential to the meaning of sentences. In this thesis, we propose to: (1) explore the methodological aspects of evaluating sign language animations with facial expressions, …

Hpcnmf: A High-Performance Toolbox For Non-Negative Matrix Factorization, Karthik Devarajan, Guoli Wang

COBRA Preprint Series

Non-negative matrix factorization (NMF) is a widely used machine learning algorithm for dimension reduction of large-scale data. It has found successful applications in a variety of fields such as computational biology, neuroscience, natural language processing, information retrieval, image processing and speech recognition. In bioinformatics, for example, it has been used to extract patterns and profiles from genomic and text-mining data as well as in protein sequence and structure analysis. While the scientific performance of NMF is very promising in dealing with high dimensional data sets and complex data structures, its computational cost is high and sometimes could be critical for …

Metadata And Linked Data In Word Sense Disambiguation, Matthew Corsmeier

Library Philosophy and Practice (e-journal)

Word Sense Disambiguation (WSD) can be assisted by taking advantage of the metadata embedded in the various ontologies, lexica, databases, etc… that exist in the Semantic Web. Automated processes that exploit the links already present in the Semantic Web can strengthen parsing of word senses by using user-contributed and semantically-linked data. These processes are only possible because of a commitment to interoperability and the creation of shared standards. This paper will review some of the most heavily used Linguistic Linked Open Data (LLOD) tools and models which show the most promise for using metadata to alleviate problems caused by polysemous …

An Empirical Study Of Semantic Similarity In Wordnet And Word2vec, Abram Handler

LSU New Orleans Theses and Dissertations

This thesis performs an empirical analysis of Word2Vec by comparing its output to WordNet, a well-known, human-curated lexical database. It finds that Word2Vec tends to uncover more of certain types of semantic relations than others -- with Word2Vec returning more hypernyms, synonomyns and hyponyms than hyponyms or holonyms. It also shows the probability that neighbors separated by a given cosine distance in Word2Vec are semantically related in WordNet. This result both adds to our understanding of the still-unknown Word2Vec and helps to benchmark new semantic tools built from word vectors.

Predicting Music Genre Preferences Based On Online Comments, Andrew J. Sinclair

Master's Theses

Communication Accommodation Theory (CAT) states that individuals adapt to each other’s communicative behaviors. This adaptation is called “convergence.” In this work we explore the convergence of writing styles of users of the online music distribution plat- form SoundCloud.com. In order to evaluate our system we created a corpus of over 38,000 comments retrieved from SoundCloud in April 2014. The corpus represents comments from 8 distinct musical genres: Classical, Electronic, Hip Hop, Jazz, Country, Metal, Folk, and World. Our corpus contains: short comments, frequent misspellings, little sentence struc- ture, hashtags, emoticons, and URLs. We adapt techniques used by researchers analyzing other …

Misheard Me Oronyminator: Using Oronyms To Validate The Correctness Of Frequency Dictionaries, Jennifer G. Hughes

Master's Theses

In the field of speech recognition, an algorithm must learn to tell the difference between "a nice rock" and "a gneiss rock". These identical-sounding phrases are called oronyms. Word frequency dictionaries are often used by speech recognition systems to help resolve phonetic sequences with more than one possible orthographic phrase interpretation, by looking up which oronym of the root phonetic sequence contains the most-common words.

Our paper demonstrates a technique used to validate word frequency dictionary values. We chose to use frequency values from the UNISYN dictionary, which tallies each word on a per-occurance basis, using a proprietary text corpus, …

Csc Senior Project: Nlpstats, Michael Mease

Computer Science and Software Engineering

Natural Language Processing has recently increased in popularity. The field of authorship analysis, specifically, uses various characteristics of text quantified by markers. NLPStats serves as a tool designed to streamline marker extraction based on user needs. A flexible query system allows for custom marker requests, adjustment of result formatting, and preprocessing options. Furthermore, an efficiently designed structure ensures that users retrieve information quickly. As a whole, NLPStats enables anyone, regardless of NLP experience, to extract important information about the text of a document.