Publications by authors named "Natalie A Deiner"

Article Synopsis
  • - The study explored the potential of large language models, specifically GPT-3.5 and GPT-4, to analyze social media posts for signs of conjunctivitis outbreaks by evaluating 12,194 related tweets across various countries from 2012 to 2023.
  • - Results showed that GPT-3.5's probabilistic assessments correlated well with evaluations from human raters and with tweet volumes in some regions, while GPT-4 demonstrated even stronger correlations.
  • - Overall, the research indicates that using AI to analyze social media could serve as an effective early warning system for identifying potential infectious disease outbreaks, particularly conjunctivitis.
View Article and Find Full Text PDF