This is my project for AIP subject at FPT University, Hanoi
Apply Attention-based model in Healthcare Representation Learning and Diagnosis Prediction
In the last decade, hospital adoption of electronic health record (EHR) systems has surged, resulting in a massive increase in the amount of digital data recorded. EHR systems save information including as demographics, diagnoses, laboratory tests and results, medications, radiographic images, clinical notes, and more for each patient visit.
While EHR adoption in hospitals and clinics has the potential to improve patient care by reducing mistakes, boosting efficiency, and enhancing treatment quality, several research have discovered secondary applications for clinical informatics. Medical concept extraction [4], [5], patient trajectory modeling, illness inference, clinical support systems, and other functions have all made use of the patient data included in EHR systems. As a result, it provides a rich data source for researchers.
There are 3 reasons why I choose this topic.
First, I’m genuinely interested in the idea that computer algorithms can improve human condition.
Second, there is a TV series about lives in hospital called “Hospital Playlist” that evoke the sense of humanity in myself, making me take advantage of informatics knowledge for meaningful projects.
Last but not least, my life recently is filled with hope, humor, and happiness from a doctor who, despite being in the front line fighting against COVID-19, still radiate ultimate optimism, listen to my on-going problems and give good advices. Therefore, this project is my attempt to understand his field better.
- Understand state-of-the-art of deep learning models and data mining methods in EHR
- Desgin, implement and optimize deep learning algorithms from scratch
- Output is a well-written documents and clean code
- One step toward AI career in healthcare domain