MEDICAL DEVICES PART II- Public Health and FDA Regulations

Welcome to the 2nd lecture! We went over the following in the last lecture: – Learned the basics of AI and ML – How is AI/ML related to biological networks – How an AI system is trained and what does it mean – Why, how, and where AI/ML helps AI is generally applied to the detection or presence of a certain disease. The manner in which it can be used to solve differs on the problem construction. In some cases, the training is done on actual images, or images are constructed from the data itself, or time series analysis are done etc. Also, depending upon the problem, it may be useful to consider what type of network should be used: CNN, DNN, RNN, LSTM etc. We will start with an understanding of some biological signals that are necessary to solve a certain detection or estimation problem. We will review the following examples: – Imaging – detection of abnormalities in a CT scan, MRI, or X-ray, – Detection of Atrial Fibrillation using: – Time series analysis – Image analysis – Drug discovery, – Help in diagnosis of a disease (automated doctor?), and – LLM in medicine How AI is used to develop an algorithm is different from how the medical devices were made in the past. AI generally requires large databases and also may or may not work on known features or aspects of a disease. In some cases, the data scientist expects AI to figure out such features. This poses a challenge to the regulating authorities that the detection of a disease is done without knowing how it was done. We will review how and what FDA is doing about it. If time permits, we will go over some cautions around data analysis. Speaker(s): Dr Md Usman Agenda: – 12:10 to 12:20 PM : Introduction by Gora Datta, FHL7 – 12:20 to 1::55 PM: Expert Lecture by Dr Md Usman – Understand some examples of using AI, – Know when to use AI for health measurements, – Regulations around AI by the FDA. – 1:55 to 2 PM : Wrap Up Room: Shark Cove, Bldg: Bealle Applied Innovation, 5270 California Ave, Irvine, California, United States, 92617, Virtual: https://events.vtools.ieee.org/m/472277