Microscopes changed how people could see and understand their surroundings. I am interested in how machine learning can be a new type of microscope to automatically discover patterns in time series data that can predict disease and model disease progression.
There is an increasing abundance of medical time series data available, such as electronic medical records (EMR) and recordings of physiological waveforms, like heart rates and respiratory rates. Machine learning can allow us to harness the information in these signals and create tools to assist clinicians and improve patient care. I am particularly fascinated by questions about how to use time series data to model the evolving state of a patient as their disease progresses and to detect deviations from the patient's normal state.