Today's health infrastructure makes it challenging for individuals to equitably access even basic medical resources. In this course, we will learn how to create modern health sensing systems that reimagine the way that healthcare delivery is performed. Specifically, we will learn how to transform ubiquitous smart devices around us like smartphones, speakers, and watches, as well as emerging wearables like earables and smartglasses, into personal medical tricorders that have the ability to provide access to health testing at our fingertips. We will learn how to tap into the rich sensor data streams (e.g. acoustic, vision, IMU) from these devices and understand core techniques in applied signal processing and machine learning to intelligently transform sensor data into clinically-relevant biomarkers which can be used to screen and diagnose diseases at scale. Beyond these techniques, this course delves into the full lifecycle of system building including ideation of frugal designs, iterative prototyping, pilot data collection, visualization, and debugging. Finally, to ensure our systems will have impact in the real-world we will cover important issues of privacy-preserving techniques and regulatory pathways which are important considerations when deploying health research. The course will focus on class discussions, hands-on demonstrations, and tutorials. Students will be evaluated on their class participation, multiple mini projects, and a final team project.
Instructors: Mayank Goel, Justin Chan (Joint office hours, TBD, Common Space outside TCS235) TA: Siqi Zhang (Office hours: TBD, Common Space outside TCS235) Location: Tepper 3500 Time: Tuesdays and Thursdays, 2:00-3:20 PM
Canvas: https://canvas.cmu.edu/courses/55703 Piazza: https://piazza.com/class/mswlpfazcjd6w1/
Assignment 0: Getting started **(Due: Sep 1)** Assignment 1: Step into smartphone sensing (Due: Sept 15) Assignment 2: Featuring engineering and activity recognition (Due: Oct 1) Assignment 3: Silent Signals (Due: Oct 20) Assignment 4: Vital Visions (Due: Nov 5) Assignment 5: Machine learning for clinical time series data (Due: Nov 24)
Summarize the paper, list pros and cons (2-4 is sufficient).
Importantly add your own subjective opinion about the work, what do you like or dislike about it? How might you build on it? How does it compare to related works at the time or today?
These reviews will contribute to the participation score.