arXiv subjects

Jay Ligatti

Publications and source records attributed to Jay Ligatti.

3 recordsLinked to original sources

arXiv2025

Topic and Sentiment Trends in Semaglutide Discussions on X: Subpopulation-Based Longitudinal Analysis

Background: User experience strongly influences pharmaceutical drug effectiveness. Social media platforms like X have become major spaces where people share medication-related experiences, especially for widely marketed drugs such as semaglutide. Despite high activity online, how different user groups engage in semaglutide discussions remains unclear. Objective: This study examines how semaglutide is perceived and discussed across X user groups by analyzing (1) changes in sentiment over time and (2) key discussion topics. Methods: We collected 859,751 posts about semaglutide from July 2021 to April 2024, along with metadata. We performed sentiment analysis and topic modeling to evaluate patterns across user subpopulations and time periods. Results: The overall mean sentiment was -0.24, with all groups showing declines over time. Discussions focused on weight loss, side effects, costs, and celebrity or political influence. Organizational accounts expressed less negative sentiment (mean = -0.04) than individuals (mean = -0.28), a statistically significant difference (P < .001). An interrupted time-series analysis showed a sentiment drop between Nov 2022 and Jan 2023, coinciding with regulatory announcements. We also found gender differences: posts by female users contained more discussions of celebrities and politicians (21 percent) compared to male users (17 percent), while male users expressed more positive sentiment. Conclusions: This study highlights how diverse user groups perceive and discuss semaglutide. Although sentiment was broadly negative, important differences emerged across subpopulations. These findings have implications for health communication and pharmacovigilance.

cs.SI↗

arXiv2024

Toward satisfactory public accessibility: A crowdsourcing approach through online reviews to inclusive urban design

As urban populations grow, the need for accessible urban design has become urgent. Traditional survey methods for assessing public perceptions of accessibility are often limited in scope. Crowdsourcing via online reviews offers a valuable alternative to understanding public perceptions, and advancements in large language models can facilitate their use. This study uses Google Maps reviews across the United States and fine-tunes Llama 3 model with the Low-Rank Adaptation technique to analyze public sentiment on accessibility. At the POI level, most categories -- restaurants, retail, hotels, and healthcare -- show negative sentiments. Socio-spatial analysis reveals that areas with higher proportions of white residents and greater socioeconomic status report more positive sentiment, while areas with more elderly, highly-educated residents exhibit more negative sentiment. Interestingly, no clear link is found between the presence of disabilities and public sentiments. Overall, this study highlights the potential of crowdsourcing for identifying accessibility challenges and providing insights for urban planners.

cs.SI↗

arXiv2024

A Preliminary Study on Using Large Language Models in Software Pentesting

Large language models (LLM) are perceived to offer promising potentials for automating security tasks, such as those found in security operation centers (SOCs). As a first step towards evaluating this perceived potential, we investigate the use of LLMs in software pentesting, where the main task is to automatically identify software security vulnerabilities in source code. We hypothesize that an LLM-based AI agent can be improved over time for a specific security task as human operators interact with it. Such improvement can be made, as a first step, by engineering prompts fed to the LLM based on the responses produced, to include relevant contexts and structures so that the model provides more accurate results. Such engineering efforts become sustainable if the prompts that are engineered to produce better results on current tasks, also produce better results on future unknown tasks. To examine this hypothesis, we utilize the OWASP Benchmark Project 1.2 which contains 2,740 hand-crafted source code test cases containing various types of vulnerabilities. We divide the test cases into training and testing data, where we engineer the prompts based on the training data (only), and evaluate the final system on the testing data. We compare the AI agent's performance on the testing data against the performance of the agent without the prompt engineering. We also compare the AI agent's results against those from SonarQube, a widely used static code analyzer for security testing. We built and tested multiple versions of the AI agent using different off-the-shelf LLMs -- Google's Gemini-pro, as well as OpenAI's GPT-3.5-Turbo and GPT-4-Turbo (with both chat completion and assistant APIs). The results show that using LLMs is a viable approach to build an AI agent for software pentesting that can improve through repeated use and prompt engineering.

cs.CR↗