2022 Spring DHC5035 Deep Learning in Medicine

 

Course Description

    1. Introduction of the more advanced models of deep learning.

      • As a consecutive class of “DHC5035 Machine Learning in Medicine“, this class will cover the remaining parts of ”Dive into Deep Learning (Freely available on https://d2l.ai/)
      • From Chapter 9 Modern RNN to Chapter 17 GAN.
      • DHC5036 Machine Learning in Medicine (2021 Fall semester) covered Chapter 1 to Chapter 8.
    2. Introduction of MI-CLAIM (Minimum Information about Clinical Artificial Intelligence Modeling) checklist & SPRIT-AI extension (Guidelines for clinical trial protocols for interventions involving artificial intelligence) checklist

      • Students can learn how to prepare clinical AI research and write the manuscript.

Instructor

    • Prof. Soo-Yong Shin (sy.shin (at) skku.edu)

TAs

    • Hyunwoo Choo (jh.choo (at) skku.edu)

Date & Room

    • Friday 9:00 AM ~ 11:50 AM
    • Room: 정약용홀, 9F, Building B, Ilwon Campus
    • On/Off hybrid class (real-time streaming at the offline classroom).
      • SAIHST students should attend the offline classroom.
      • Other students (including part-time students) could attend the class via webex.
      • Online link: https://skku-ict.webex.com/meet/sy.shin 
      • After class, the recorded lecture will be uploaded in iCampus. Part-time students can watch this recorded lecture.

Textbook

    • PyTorch
    • Python, Numpy and other necessary libraries for deep learning
    • Kaggle Notebook (based on Jupyter notebook)
    • Bring your own laptop for programming practice

Evaluation

    • Attendance: P/F
    • Assignment: 30%
    • Paper Reading & Presentation: 30%
    • Term Project
      • Initial Presentation: 10%
      • Final Presentation: 30%
    • FINAL GRADE

Assignments

    • All assignments should be submitted to GitHub (https://classroom.github.com/a/3UqVVPV8)
    • How to submit your assignments (If you need assistant, contact TA)
    • 6 Assignments
    • Students should submit an assignment before next class (Thursday midnight). If late, the score will be ZERO.
    • File naming convention:  Assignment_1(change the number based on the assignment)_StudentID_StudentName.extension

Paper Reading

Term Project

    • The student should present two times for the term project.
      • Initial presentation: The student should present the chosen topic 
      • Final presentation: The student should present the (tentative) experimental results. 
      • All presentation should be based on MI-CLAIM checklist & SPRIT-AI extension.
    • Each presentation should be more 30 minutes (no maximum as well).

Course Schedule

(Course schedule can be changed.)

    1. 2/25: Course Introduction
    2. 3/4: Modern RNN & Practice
    3. 3/11:  Attention Mechanism & Practice1, Practice2, Practice3
      • Assignment #2 – (Score): Assignment is given in the end of practice 3.
    4. 3/18: Optimization Algorithm & Practice
      • Assignment #3 – (Score)
    5. 3/25: Computer Vision & Practice 1, Practice 2, Practice 3, Practice 4, Examples (Practice 1 & 3 were updated)
      • Assignment #4 – (Score)
    6. 4/1: National Language Processing (NLP) & Practice 1 (with the assignment), Practice 2
      • Assignment #5 – (Score)
    7. 4/8: Generative Adversarial Networks & Practice
      • Assignment #6 – (Score)
    8. 4/15: Introduction of Term Project & Paper Reading
    9. 4/22: Term Project: Initial Presentation – (Score)
    10. 4/29 : Paper Reading & Presentation (by Students) – (Score)
    11. 5/6: Paper Reading & Presentation (by Students) – (Score
    12. 5/13: Paper Reading & Presentation (by Students) – (Score)
    13. 5/20: Term Project: Final Presentation – (Score)
    14. 5/27: Term Project: Final Presentation – (Score)
    15. 6/3: Term Project: Final Presentation – (Score)
  1.