Abstract:
The goal of this work is to develop a system that can diagnose and identify patients using AI and
machine learning. Shortness of breath and cough are among the most common, urgent, and lifethreatening conditions in emergency departments. Cough holds critical information about many
respiratory diseases and is a symptom of more than twenty different conditions. The hidden features
within a cough can also be used for early disease detection by leveraging trained AI algorithms that
analyze cough acoustic data. Approximately half of adults experiencing shortness of breath suffer from
acute heart failure (AHF), exacerbation of chronic obstructive pulmonary disease (COPD), or
pneumonia. These conditions are often misdiagnosed and, as a result, poorly treated in emergency care.
Our goal is to develop AI-based diagnostic decision support with interpretable and individualized
features, utilizing data collected from across the region.
