6.871/HST.956: Machine Learning for Healthcare

Instructors: David Sontag, Madhur Nayan
Teaching Assistants: Zeshan Hussain, Intae Moon
Graduate level; Units 4-0-8 (counts as an AUS2, AAGS, and II subject; also a EECS AI TQE)
Time: Tuesdays & Thursdays, 9:30-11:00am EST
Location: 32-141
Prerequisite: 6.036/6.862 or 6.867 or 9.520/6.860 or 6.806/6.864 or 6.438 or 6.034
Office Hours: Monday 4:00-5:00pm (Room 26-168), Friday 4:00-5:00pm (Room 36-112)
Recitations (required): 4-270, Friday 3:00-4:00pm EST
Contact staffs: mlhc22mit@gmail.com

Announcements

The final project poster session will be held at the Grier room (34-401) from 9:30am-12pm on May 5th.
Pset 5 has been posted. It's due Apr. 20th (Wed), 11:59 PM ET.
Pset 4 has been posted. It's due Mar. 18th (Fri), 11:59 PM ET.
Pset 3 has been posted. It's due Mar. 9th (Wed), 11:59 PM ET.
Project guidelines have been posted; Project checkpoint 1 is due Mar. 3rd (Thu), 11:59 PM ET. Please visit Canvas page for detail.
Pset 2 has been posted. It's due Feb. 25th (Fri), 11:59 PM ET.
Pset 1 has been posted. It's due Feb. 16th (Wed), 11:59 PM ET.
Pset 0 has been posted. It's due Feb. 2nd (Wed), 11:59 PM ET.
We signed the class up for MIT's "lightweight lecture capture pilot" which should provide the ability to record lectures in 32-141. However, we do not plan to support remote participation in lectures and our expectation is that students will attend class in person unless an exception is granted (e.g., students needing to quarantine/isolate, or a clinician who is on service).
See 2021's site for last year's schedule and problem sets; this year's will be similar. See 2020's site or MIT OpenCourseWare for slides and videos of past lectures.
This class will have a final exam, but no midterm.
Recitation this year will be required (units have been changed in MIT catalog to account for this); they'll include deep dives into medical data and paper discussions.

Course description

Introduces students to machine learning in healthcare, including the nature of clinical data and the use of machine learning for risk stratification, disease progression modeling, precision medicine, diagnosis, subtype discovery, and improving clinical workflows. Topics include causality, interpretability, algorithmic fairness, time-series analysis, graphical models, deep learning and transfer learning. Guest lectures by clinicians from the Boston area and course projects with real clinical data emphasize subtleties of working with clinical data and translating machine learning into clinical practice.

Schedule

Please visit Canvas page to access all slide decks, recitation materials, and more.

Class Date Lecture Materials Assignments

Unit 1: Overview of Clinical Care & Data

1 Tuesday Feb 1
Introduction: What makes healthcare unique? [slides]
Week 1 reading :

[required] AI in health and medicine

[optional] Machine Learning in Medicine

Project preference form

Reading response form

Pset 0 (due Feb 2)

2 Thursday Feb 3
Overview of Clinical Care [slides]

3 Tuesday Feb 8
Overview of Clinical Data Science [slides]

Pset 1(due Feb 16)

Unit 2: ML with Clinical Text, Imaging, Physiological, and Administrative Data

4 Thursday Feb 10
Risk Stratification from Structured Health Data [slides]

5 Tuesday Feb 15
Physiological Time-Series [slides]

Pset 2(due Feb 25)

6 Thursday Feb 17
Intro to Clinical NLP (Monica Agrawal) [slides]

Tuesday Feb 22
Monday Schedule - No Class

7 Thursday Feb 24
Clinical NLP for oncology & application to AI-driven documentation &
(Guest speakers: Kenneth Kehl, DFCI & Monica Agrawal, MIT)

Pset 3(due Mar 9)

8 Tuesday Mar 1
Intro to ML for medical imaging: algorithms &
application to mammograpy (Adam Yala) [slides]

9 Thursday Mar 3
ML for pathology (Guest speaker: Andy Beck, MD/PhD (CEO, PathAI))

Unit 3: Causal Inference

10 Tuesday Mar 8
Causal graphs, potential outcomes, covariate adjustment [slides]

11 Thursday Mar 10
Assumptions for causal inference, inverse propensity weighting [slides]

Pset 4(due Mar 18)

12 Tuesday Mar 15
Policy learning, sensitivity analysis, instrumental variables [slides]

13 Thursday Mar 17
Dataset shift [slides]

Spring Break (Mar 21 - 25)

Unit 4: Methods for ML in Healthcare

14 Tuesday Mar 29
Critical appraisal of applying ML in healthcare [slides]
reading :

[required] Protocol for development of a reporting guideline (TRIPOD-AI) and risk of bias tool (PROBAST-AI)...

Reading response form

15 Thursday Mar 31
Weak supervision & learning from noisy labels [slides]

16 Tuesday Apr 5
Survival modeling [slides]

[optional] Survival Analysis: A Self-Learning Text

[optional] Survival Analysis: Techniques for Censored and Truncated Data

Pset 5(due Apr 20)

17 Thursday Apr 7
Fairness (Irene Chen) [slides]

[required] Dissecting racial bias in an algorithm used to manage the health of populations

[optional] Fairness And Machine Learning : Limitations and Opportunities

Reading response form

18 Tuesday Apr 12
Interpretability [slides]

[optional] Interpretable Machine Learning : A Guide for Making Black Box Models Explainable

19 Thursday Apr 14
Human-AI interaction (Hussein Mozannar) [slides]

20 Tuesday Apr 19
Disease subtyping & progression modeling [slides]

21 Thursday Apr 21
Privacy [slides] + Guest speaker: Fei Wang, Weill Cornell Medicine [slides]

Unit 5: Off-policy reinforcement learning & Regulation

22 Tuesday Apr 26
Intro to off-policy reinforcement learning and dynamic treatment regimes [slides]

23 Thursday Apr 28
Guest lecture: Susan Murphy (Harvard)

24 Tuesday May 3
Regulation of AI in healthcare

[optional] Butterfly rescinds features from handheld ultrasound device, working with FDA to re-enable

Project poster session & final exam review

25 Thursday May 5
Project poster session, 9:30am-12pm at the Grier room (34-401)

26 Tuesday May 10
Final exam review

Class	Date	Lecture Materials	Assignments
Unit 1: Overview of Clinical Care & Data
1	Tuesday Feb 1	Introduction: What makes healthcare unique? [slides] Week 1 reading : [required] AI in health and medicine [optional] Machine Learning in Medicine	Project preference form Reading response form Pset 0 (due Feb 2)
2	Thursday Feb 3	Overview of Clinical Care [slides]
3	Tuesday Feb 8	Overview of Clinical Data Science [slides]	Pset 1(due Feb 16)
Unit 2: ML with Clinical Text, Imaging, Physiological, and Administrative Data
4	Thursday Feb 10	Risk Stratification from Structured Health Data [slides]
5	Tuesday Feb 15	Physiological Time-Series [slides]	Pset 2(due Feb 25)
6	Thursday Feb 17	Intro to Clinical NLP (Monica Agrawal) [slides]
	Tuesday Feb 22	Monday Schedule - No Class
7	Thursday Feb 24	Clinical NLP for oncology & application to AI-driven documentation & (Guest speakers: Kenneth Kehl, DFCI & Monica Agrawal, MIT)	Pset 3(due Mar 9)
8	Tuesday Mar 1	Intro to ML for medical imaging: algorithms & application to mammograpy (Adam Yala) [slides]
9	Thursday Mar 3	ML for pathology (Guest speaker: Andy Beck, MD/PhD (CEO, PathAI))
Unit 3: Causal Inference
10	Tuesday Mar 8	Causal graphs, potential outcomes, covariate adjustment [slides]
11	Thursday Mar 10	Assumptions for causal inference, inverse propensity weighting [slides]	Pset 4(due Mar 18)
12	Tuesday Mar 15	Policy learning, sensitivity analysis, instrumental variables [slides]
13	Thursday Mar 17	Dataset shift [slides]
Spring Break (Mar 21 - 25)
Unit 4: Methods for ML in Healthcare
14	Tuesday Mar 29	Critical appraisal of applying ML in healthcare [slides] reading : [required] Protocol for development of a reporting guideline (TRIPOD-AI) and risk of bias tool (PROBAST-AI)...	Reading response form
15	Thursday Mar 31	Weak supervision & learning from noisy labels [slides]
16	Tuesday Apr 5	Survival modeling [slides] [optional] Survival Analysis: A Self-Learning Text [optional] Survival Analysis: Techniques for Censored and Truncated Data	Pset 5(due Apr 20)
17	Thursday Apr 7	Fairness (Irene Chen) [slides] [required] Dissecting racial bias in an algorithm used to manage the health of populations [optional] Fairness And Machine Learning : Limitations and Opportunities	Reading response form
18	Tuesday Apr 12	Interpretability [slides] [optional] Interpretable Machine Learning : A Guide for Making Black Box Models Explainable
19	Thursday Apr 14	Human-AI interaction (Hussein Mozannar) [slides]
20	Tuesday Apr 19	Disease subtyping & progression modeling [slides]
21	Thursday Apr 21	Privacy [slides] + Guest speaker: Fei Wang, Weill Cornell Medicine [slides]
Unit 5: Off-policy reinforcement learning & Regulation
22	Tuesday Apr 26	Intro to off-policy reinforcement learning and dynamic treatment regimes [slides]
23	Thursday Apr 28	Guest lecture: Susan Murphy (Harvard)
24	Tuesday May 3	Regulation of AI in healthcare [optional] Butterfly rescinds features from handheld ultrasound device, working with FDA to re-enable
Project poster session & final exam review
25	Thursday May 5	Project poster session, 9:30am-12pm at the Grier room (34-401)
26	Tuesday May 10	Final exam review

Problem sets

Pset 0 : in order to access sensitive healthcare datasets, you will need to complete several preliminary tasks, including the CITI training and filling out the Data Use Agreement of MIMIC-III dataset. Please upload the full CITI completion report to Canvas once you complete the requested tasks.
Pset 1
Pset 2
Pset 3
Pset 4
Pset 5

Projects

We will release the project guidelines and detailed information in Canvas.

Late Policy

(starting for pset1 onwards)

[2 "slack" days] We understand that sometimes things outside one's control prevent submitting by the deadline. As such, each student is given 2 "slack" days that they can use throughout the semester (e.g. you could submit two psets one day late each or you could submit one pset two days late) without a late penalty. the days do not subdivide into sub-day units: 2 hours late would spend one of the slack days without 22 hours of "rollover". In your pdf writeup, specify how many slack days you are using (they cannot be used retroactively).
[10% off per unexcused late day.] If you submit a pset 3 days late and use 1 slack day, then this is 2 unexcused late days, which translates to 20% off your homework.
[Max 4 days late.] Homework will not be accepted after 4 days late, regardless of how many slack days are used, absent communication from S3 or OGE.
[write on homework] In order to use a slack day, students must include it in writing on their submission pdf. Otherwise, TAs will assume no slack days used and deduct 10% for each late day.

Scenarios:

Sam uses 2 slack days on HW3. This is the first time Sam has used any slack days. Sam now has 0 remaining slack days and receives her homework score with no penalty.
Jamie uses 1 slack day on HW3 but submits 52 hours after the deadline. Therefore Jamie is 3 days late (rounded up) and receives 20% off the graded homework. This is the first time Jamie has used any slack days, so Jamie now has 1 slack day remaining.
Cory has never used any slack days, but submits HW5 100 hours late with no note / communication from S3 or OGE. This is more than 4 days after the deadline, so the work is not accepted, and Cory receives a zero.